ATATC commited on
Commit
dbfee17
·
verified ·
1 Parent(s): 77ce13f

Upload folder using huggingface_hub

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +8 -0
  2. flare-medgemma-base-eval/details.json +1113 -0
  3. flare-medgemma-base-eval/scores.json +214 -0
  4. flare-medgemma-base-infer/inference_details.json +40 -0
  5. flare-medgemma-base-infer/testing_predictions.jsonl +0 -0
  6. flare-medgemma-base-infer/validation_hidden_predictions.jsonl +0 -0
  7. flare-medgemma-base-infer/validation_public_predictions.jsonl +0 -0
  8. flare-medgemma-eval/details.json +1113 -0
  9. flare-medgemma-eval/scores.json +214 -0
  10. flare-medgemma-infer/inference_details.json +40 -0
  11. flare-medgemma-infer/testing_predictions.jsonl +0 -0
  12. flare-medgemma-infer/validation_hidden_predictions.jsonl +0 -0
  13. flare-medgemma-infer/validation_public_predictions.jsonl +0 -0
  14. flare-medgemma-medgemma15-lora/README.md +58 -0
  15. flare-medgemma-medgemma15-lora/checkpoint-17000/README.md +209 -0
  16. flare-medgemma-medgemma15-lora/checkpoint-17000/adapter_config.json +51 -0
  17. flare-medgemma-medgemma15-lora/checkpoint-17000/adapter_model.safetensors +3 -0
  18. flare-medgemma-medgemma15-lora/checkpoint-17000/chat_template.jinja +47 -0
  19. flare-medgemma-medgemma15-lora/checkpoint-17000/optimizer.pt +3 -0
  20. flare-medgemma-medgemma15-lora/checkpoint-17000/processor_config.json +28 -0
  21. flare-medgemma-medgemma15-lora/checkpoint-17000/rng_state.pth +3 -0
  22. flare-medgemma-medgemma15-lora/checkpoint-17000/scheduler.pt +3 -0
  23. flare-medgemma-medgemma15-lora/checkpoint-17000/tokenizer.json +3 -0
  24. flare-medgemma-medgemma15-lora/checkpoint-17000/tokenizer_config.json +26 -0
  25. flare-medgemma-medgemma15-lora/checkpoint-17000/trainer_state.json +0 -0
  26. flare-medgemma-medgemma15-lora/checkpoint-17000/training_args.bin +3 -0
  27. flare-medgemma-medgemma15-lora/checkpoint-17200/README.md +209 -0
  28. flare-medgemma-medgemma15-lora/checkpoint-17200/adapter_config.json +51 -0
  29. flare-medgemma-medgemma15-lora/checkpoint-17200/adapter_model.safetensors +3 -0
  30. flare-medgemma-medgemma15-lora/checkpoint-17200/chat_template.jinja +47 -0
  31. flare-medgemma-medgemma15-lora/checkpoint-17200/optimizer.pt +3 -0
  32. flare-medgemma-medgemma15-lora/checkpoint-17200/processor_config.json +28 -0
  33. flare-medgemma-medgemma15-lora/checkpoint-17200/rng_state.pth +3 -0
  34. flare-medgemma-medgemma15-lora/checkpoint-17200/scheduler.pt +3 -0
  35. flare-medgemma-medgemma15-lora/checkpoint-17200/tokenizer.json +3 -0
  36. flare-medgemma-medgemma15-lora/checkpoint-17200/tokenizer_config.json +26 -0
  37. flare-medgemma-medgemma15-lora/checkpoint-17200/trainer_state.json +0 -0
  38. flare-medgemma-medgemma15-lora/checkpoint-17200/training_args.bin +3 -0
  39. flare-medgemma-medgemma15-lora/checkpoint-17208/README.md +209 -0
  40. flare-medgemma-medgemma15-lora/checkpoint-17208/adapter_config.json +51 -0
  41. flare-medgemma-medgemma15-lora/checkpoint-17208/adapter_model.safetensors +3 -0
  42. flare-medgemma-medgemma15-lora/checkpoint-17208/chat_template.jinja +47 -0
  43. flare-medgemma-medgemma15-lora/checkpoint-17208/optimizer.pt +3 -0
  44. flare-medgemma-medgemma15-lora/checkpoint-17208/processor_config.json +28 -0
  45. flare-medgemma-medgemma15-lora/checkpoint-17208/rng_state.pth +3 -0
  46. flare-medgemma-medgemma15-lora/checkpoint-17208/scheduler.pt +3 -0
  47. flare-medgemma-medgemma15-lora/checkpoint-17208/tokenizer.json +3 -0
  48. flare-medgemma-medgemma15-lora/checkpoint-17208/tokenizer_config.json +26 -0
  49. flare-medgemma-medgemma15-lora/checkpoint-17208/trainer_state.json +0 -0
  50. flare-medgemma-medgemma15-lora/checkpoint-17208/training_args.bin +3 -0
.gitattributes CHANGED
@@ -33,3 +33,11 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
 
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ flare-medgemma-medgemma15-lora/checkpoint-17000/tokenizer.json filter=lfs diff=lfs merge=lfs -text
37
+ flare-medgemma-medgemma15-lora/checkpoint-17200/tokenizer.json filter=lfs diff=lfs merge=lfs -text
38
+ flare-medgemma-medgemma15-lora/checkpoint-17208/tokenizer.json filter=lfs diff=lfs merge=lfs -text
39
+ flare-medgemma-medgemma15-lora/final/tokenizer.json filter=lfs diff=lfs merge=lfs -text
40
+ flare-medgemma1-medgemma1-lora/checkpoint-8400/tokenizer.json filter=lfs diff=lfs merge=lfs -text
41
+ flare-medgemma1-medgemma1-lora/checkpoint-8600/tokenizer.json filter=lfs diff=lfs merge=lfs -text
42
+ flare-medgemma1-medgemma1-lora/checkpoint-8604/tokenizer.json filter=lfs diff=lfs merge=lfs -text
43
+ flare-medgemma1-medgemma1-lora/final/tokenizer.json filter=lfs diff=lfs merge=lfs -text
flare-medgemma-base-eval/details.json ADDED
@@ -0,0 +1,1113 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metrics": {
3
+ "balanced_accuracy": 0.015349964696408524,
4
+ "cell_counting_mean_absolute_error": 32663.9,
5
+ "cell_counting_root_mean_squared_error": 173393.05504108287,
6
+ "classification_accuracy": 0.0275212196533894,
7
+ "classification_f1_score": 0.03163966776383068,
8
+ "crimson_score": 0.69777170781893,
9
+ "detection_average_f1": 0.004551652647125287,
10
+ "detection_average_precision": 0.00340978076326286,
11
+ "detection_average_recall": 0.013534758655479323,
12
+ "detection_f1_iou_0.3": 0.015570961364704957,
13
+ "detection_f1_iou_0.5": 0.0007284469659534537,
14
+ "detection_iou_0.3_f1_score": 0.015570961364704957,
15
+ "detection_iou_0.3_fn": 8494.0,
16
+ "detection_iou_0.3_fp": 2703.6666666666665,
17
+ "detection_iou_0.3_precision": 0.011692169169451113,
18
+ "detection_iou_0.3_recall": 0.04630080803305465,
19
+ "detection_iou_0.3_tp": 22.0,
20
+ "detection_iou_0.4_f1_score": 0.00639943455052561,
21
+ "detection_iou_0.4_fn": 8508.0,
22
+ "detection_iou_0.4_fp": 2717.6666666666665,
23
+ "detection_iou_0.4_precision": 0.004477958626839689,
24
+ "detection_iou_0.4_recall": 0.019470639027858148,
25
+ "detection_iou_0.4_tp": 8.0,
26
+ "detection_iou_0.5_f1_score": 0.0007284469659534537,
27
+ "detection_iou_0.5_fn": 8514.0,
28
+ "detection_iou_0.5_fp": 2723.6666666666665,
29
+ "detection_iou_0.5_precision": 0.0007515495569191695,
30
+ "detection_iou_0.5_recall": 0.0018635842729599608,
31
+ "detection_iou_0.5_tp": 2.0,
32
+ "detection_iou_0.6_f1_score": 2.971017722120712e-05,
33
+ "detection_iou_0.6_fn": 8515.666666666666,
34
+ "detection_iou_0.6_fp": 2725.3333333333335,
35
+ "detection_iou_0.6_precision": 6.361323155216285e-05,
36
+ "detection_iou_0.6_recall": 1.938097176192414e-05,
37
+ "detection_iou_0.6_tp": 0.3333333333333333,
38
+ "detection_iou_0.7_f1_score": 2.971017722120712e-05,
39
+ "detection_iou_0.7_fn": 8515.666666666666,
40
+ "detection_iou_0.7_fp": 2725.3333333333335,
41
+ "detection_iou_0.7_precision": 6.361323155216285e-05,
42
+ "detection_iou_0.7_recall": 1.938097176192414e-05,
43
+ "detection_iou_0.7_tp": 0.3333333333333333,
44
+ "detection_precision_iou_0.3": 0.011692169169451113,
45
+ "detection_precision_iou_0.5": 0.0007515495569191695,
46
+ "detection_recall_iou_0.3": 0.04630080803305465,
47
+ "detection_recall_iou_0.5": 0.0018635842729599608,
48
+ "f1_iou_0.5": 0.0007284469659534537,
49
+ "f1_score": 0.15408583987692842,
50
+ "green_score": 0.771889052107561,
51
+ "multi_label_f1_score": 0.15408583987692842,
52
+ "multi_label_precision": 0.10280958512649903,
53
+ "multi_label_recall": 0.4823593453229044,
54
+ "regression_mean_absolute_error": 28.20773801553721,
55
+ "regression_root_mean_squared_error": 40.32094855174149
56
+ },
57
+ "mean_metrics": {
58
+ "balanced_accuracy": 0.015349964696408524,
59
+ "cell_counting_mean_absolute_error": 32663.9,
60
+ "cell_counting_root_mean_squared_error": 173393.05504108287,
61
+ "classification_accuracy": 0.0275212196533894,
62
+ "classification_f1_score": 0.03163966776383068,
63
+ "crimson_score": 0.69777170781893,
64
+ "detection_average_f1": 0.004551652647125287,
65
+ "detection_average_precision": 0.00340978076326286,
66
+ "detection_average_recall": 0.013534758655479323,
67
+ "detection_f1_iou_0.3": 0.015570961364704957,
68
+ "detection_f1_iou_0.5": 0.0007284469659534537,
69
+ "detection_iou_0.3_f1_score": 0.015570961364704957,
70
+ "detection_iou_0.3_fn": 8494.0,
71
+ "detection_iou_0.3_fp": 2703.6666666666665,
72
+ "detection_iou_0.3_precision": 0.011692169169451113,
73
+ "detection_iou_0.3_recall": 0.04630080803305465,
74
+ "detection_iou_0.3_tp": 22.0,
75
+ "detection_iou_0.4_f1_score": 0.00639943455052561,
76
+ "detection_iou_0.4_fn": 8508.0,
77
+ "detection_iou_0.4_fp": 2717.6666666666665,
78
+ "detection_iou_0.4_precision": 0.004477958626839689,
79
+ "detection_iou_0.4_recall": 0.019470639027858148,
80
+ "detection_iou_0.4_tp": 8.0,
81
+ "detection_iou_0.5_f1_score": 0.0007284469659534537,
82
+ "detection_iou_0.5_fn": 8514.0,
83
+ "detection_iou_0.5_fp": 2723.6666666666665,
84
+ "detection_iou_0.5_precision": 0.0007515495569191695,
85
+ "detection_iou_0.5_recall": 0.0018635842729599608,
86
+ "detection_iou_0.5_tp": 2.0,
87
+ "detection_iou_0.6_f1_score": 2.971017722120712e-05,
88
+ "detection_iou_0.6_fn": 8515.666666666666,
89
+ "detection_iou_0.6_fp": 2725.3333333333335,
90
+ "detection_iou_0.6_precision": 6.361323155216285e-05,
91
+ "detection_iou_0.6_recall": 1.938097176192414e-05,
92
+ "detection_iou_0.6_tp": 0.3333333333333333,
93
+ "detection_iou_0.7_f1_score": 2.971017722120712e-05,
94
+ "detection_iou_0.7_fn": 8515.666666666666,
95
+ "detection_iou_0.7_fp": 2725.3333333333335,
96
+ "detection_iou_0.7_precision": 6.361323155216285e-05,
97
+ "detection_iou_0.7_recall": 1.938097176192414e-05,
98
+ "detection_iou_0.7_tp": 0.3333333333333333,
99
+ "detection_precision_iou_0.3": 0.011692169169451113,
100
+ "detection_precision_iou_0.5": 0.0007515495569191695,
101
+ "detection_recall_iou_0.3": 0.04630080803305465,
102
+ "detection_recall_iou_0.5": 0.0018635842729599608,
103
+ "f1_iou_0.5": 0.0007284469659534537,
104
+ "f1_score": 0.15408583987692842,
105
+ "green_score": 0.771889052107561,
106
+ "multi_label_f1_score": 0.15408583987692842,
107
+ "multi_label_precision": 0.10280958512649903,
108
+ "multi_label_recall": 0.4823593453229044,
109
+ "regression_mean_absolute_error": 28.20773801553721,
110
+ "regression_root_mean_squared_error": 40.32094855174149
111
+ },
112
+ "by_task": {
113
+ "cell_counting": {
114
+ "metric": "mean_absolute_error",
115
+ "value": 32663.9,
116
+ "count": 100,
117
+ "per_split": {
118
+ "validation_public": 32663.9
119
+ }
120
+ },
121
+ "detection": {
122
+ "metric": "f1_iou_0.5",
123
+ "value": 0.0007284469659534537,
124
+ "count": 1392,
125
+ "per_split": {
126
+ "testing": 0.0003565221266544854,
127
+ "validation_public": 0.00162999185004075,
128
+ "validation_hidden": 0.00019882692116512577
129
+ }
130
+ },
131
+ "disease_diagnosis_classification": {
132
+ "metric": "balanced_accuracy",
133
+ "value": 0.015349964696408524,
134
+ "count": 6123,
135
+ "per_split": {
136
+ "testing": 0.0020100502512562816,
137
+ "validation_public": 0.041802081600207056,
138
+ "validation_hidden": 0.0022377622377622378
139
+ }
140
+ },
141
+ "multi_label_classification": {
142
+ "metric": "f1_score",
143
+ "value": 0.15408583987692842,
144
+ "count": 2370,
145
+ "per_split": {
146
+ "testing": 0.1507593198931337,
147
+ "validation_public": 0.17710801631592055,
148
+ "validation_hidden": 0.134390183421731
149
+ }
150
+ },
151
+ "regression": {
152
+ "metric": "mean_absolute_error",
153
+ "value": 28.20773801553721,
154
+ "count": 302,
155
+ "per_split": {
156
+ "testing": 30.038603799104408,
157
+ "validation_hidden": 26.376872231970015
158
+ }
159
+ },
160
+ "report_generation": {
161
+ "metric": "green_score",
162
+ "value": 0.771889052107561,
163
+ "count": 1945,
164
+ "per_split": {
165
+ "validation_public": 0.771889052107561
166
+ }
167
+ }
168
+ },
169
+ "split_results": {
170
+ "testing": {
171
+ "metrics": {
172
+ "f1_iou_0.5": 0.0003565221266544854,
173
+ "detection_f1_iou_0.5": 0.0003565221266544854,
174
+ "detection_precision_iou_0.5": 0.0007633587786259542,
175
+ "detection_recall_iou_0.5": 0.0002325716611430897,
176
+ "detection_f1_iou_0.3": 0.0029413075448995055,
177
+ "detection_precision_iou_0.3": 0.006297709923664122,
178
+ "detection_recall_iou_0.3": 0.0019187162044304901,
179
+ "detection_average_f1": 0.000909131422968938,
180
+ "detection_average_precision": 0.0019465648854961833,
181
+ "detection_average_recall": 0.0005930577359148788,
182
+ "detection_iou_0.3_f1_score": 0.0029413075448995055,
183
+ "detection_iou_0.3_precision": 0.006297709923664122,
184
+ "detection_iou_0.3_recall": 0.0019187162044304901,
185
+ "detection_iou_0.3_tp": 33,
186
+ "detection_iou_0.3_fp": 5207,
187
+ "detection_iou_0.3_fn": 17166,
188
+ "detection_iou_0.4_f1_score": 0.0010695663799634564,
189
+ "detection_iou_0.4_precision": 0.0022900763358778627,
190
+ "detection_iou_0.4_recall": 0.0006977149834292691,
191
+ "detection_iou_0.4_tp": 12,
192
+ "detection_iou_0.4_fp": 5228,
193
+ "detection_iou_0.4_fn": 17187,
194
+ "detection_iou_0.5_f1_score": 0.0003565221266544854,
195
+ "detection_iou_0.5_precision": 0.0007633587786259542,
196
+ "detection_iou_0.5_recall": 0.0002325716611430897,
197
+ "detection_iou_0.5_tp": 4,
198
+ "detection_iou_0.5_fp": 5236,
199
+ "detection_iou_0.5_fn": 17195,
200
+ "detection_iou_0.6_f1_score": 8.913053166362135e-05,
201
+ "detection_iou_0.6_precision": 0.00019083969465648855,
202
+ "detection_iou_0.6_recall": 5.8142915285772426e-05,
203
+ "detection_iou_0.6_tp": 1,
204
+ "detection_iou_0.6_fp": 5239,
205
+ "detection_iou_0.6_fn": 17198,
206
+ "detection_iou_0.7_f1_score": 8.913053166362135e-05,
207
+ "detection_iou_0.7_precision": 0.00019083969465648855,
208
+ "detection_iou_0.7_recall": 5.8142915285772426e-05,
209
+ "detection_iou_0.7_tp": 1,
210
+ "detection_iou_0.7_fp": 5239,
211
+ "detection_iou_0.7_fn": 17198,
212
+ "balanced_accuracy": 0.0020100502512562816,
213
+ "classification_accuracy": 0.00842911877394636,
214
+ "classification_f1_score": 0.015474120258443963,
215
+ "f1_score": 0.1507593198931337,
216
+ "multi_label_f1_score": 0.1507593198931337,
217
+ "multi_label_precision": 0.10261288282875378,
218
+ "multi_label_recall": 0.40240158902130735,
219
+ "regression_mean_absolute_error": 30.038603799104408,
220
+ "regression_root_mean_squared_error": 46.64166890561044
221
+ },
222
+ "by_task": {
223
+ "detection": {
224
+ "metric": "f1_iou_0.5",
225
+ "value": 0.0003565221266544854,
226
+ "count": 961,
227
+ "metrics": {
228
+ "f1_score": 0.0003565221266544854,
229
+ "precision": 0.0007633587786259542,
230
+ "recall": 0.0002325716611430897,
231
+ "f1_score_at_03": 0.0029413075448995055,
232
+ "f1_score_at_05": 0.0003565221266544854,
233
+ "average_f1": 0.000909131422968938,
234
+ "precision_at_03": 0.006297709923664122,
235
+ "recall_at_03": 0.0019187162044304901,
236
+ "detailed_metrics": {
237
+ "IoU_0.3": {
238
+ "f1_score": 0.0029413075448995055,
239
+ "precision": 0.006297709923664122,
240
+ "recall": 0.0019187162044304901,
241
+ "tp": 33,
242
+ "fp": 5207,
243
+ "fn": 17166
244
+ },
245
+ "IoU_0.4": {
246
+ "f1_score": 0.0010695663799634564,
247
+ "precision": 0.0022900763358778627,
248
+ "recall": 0.0006977149834292691,
249
+ "tp": 12,
250
+ "fp": 5228,
251
+ "fn": 17187
252
+ },
253
+ "IoU_0.5": {
254
+ "f1_score": 0.0003565221266544854,
255
+ "precision": 0.0007633587786259542,
256
+ "recall": 0.0002325716611430897,
257
+ "tp": 4,
258
+ "fp": 5236,
259
+ "fn": 17195
260
+ },
261
+ "IoU_0.6": {
262
+ "f1_score": 8.913053166362135e-05,
263
+ "precision": 0.00019083969465648855,
264
+ "recall": 5.8142915285772426e-05,
265
+ "tp": 1,
266
+ "fp": 5239,
267
+ "fn": 17198
268
+ },
269
+ "IoU_0.7": {
270
+ "f1_score": 8.913053166362135e-05,
271
+ "precision": 0.00019083969465648855,
272
+ "recall": 5.8142915285772426e-05,
273
+ "tp": 1,
274
+ "fp": 5239,
275
+ "fn": 17198
276
+ }
277
+ },
278
+ "coco_style_metrics": {
279
+ "average_f1": 0.000909131422968938,
280
+ "average_precision": 0.0019465648854961833,
281
+ "average_recall": 0.0005930577359148788
282
+ },
283
+ "per_chromosome_metrics": {
284
+ "1": {
285
+ "f1_score": 0.0,
286
+ "precision": 0.0,
287
+ "recall": 0.0,
288
+ "tp": 0,
289
+ "fp": 0,
290
+ "fn": 725
291
+ },
292
+ "10": {
293
+ "f1_score": 0.0,
294
+ "precision": 0.0,
295
+ "recall": 0.0,
296
+ "tp": 0,
297
+ "fp": 0,
298
+ "fn": 722
299
+ },
300
+ "11": {
301
+ "f1_score": 0.0,
302
+ "precision": 0.0,
303
+ "recall": 0.0,
304
+ "tp": 0,
305
+ "fp": 0,
306
+ "fn": 723
307
+ },
308
+ "12": {
309
+ "f1_score": 0.0,
310
+ "precision": 0.0,
311
+ "recall": 0.0,
312
+ "tp": 0,
313
+ "fp": 0,
314
+ "fn": 725
315
+ },
316
+ "13": {
317
+ "f1_score": 0.0,
318
+ "precision": 0.0,
319
+ "recall": 0.0,
320
+ "tp": 0,
321
+ "fp": 0,
322
+ "fn": 721
323
+ },
324
+ "14": {
325
+ "f1_score": 0.0,
326
+ "precision": 0.0,
327
+ "recall": 0.0,
328
+ "tp": 0,
329
+ "fp": 0,
330
+ "fn": 694
331
+ },
332
+ "15": {
333
+ "f1_score": 0.0,
334
+ "precision": 0.0,
335
+ "recall": 0.0,
336
+ "tp": 0,
337
+ "fp": 0,
338
+ "fn": 714
339
+ },
340
+ "16": {
341
+ "f1_score": 0.0,
342
+ "precision": 0.0,
343
+ "recall": 0.0,
344
+ "tp": 0,
345
+ "fp": 0,
346
+ "fn": 725
347
+ },
348
+ "17": {
349
+ "f1_score": 0.0,
350
+ "precision": 0.0,
351
+ "recall": 0.0,
352
+ "tp": 0,
353
+ "fp": 0,
354
+ "fn": 723
355
+ },
356
+ "18": {
357
+ "f1_score": 0.0,
358
+ "precision": 0.0,
359
+ "recall": 0.0,
360
+ "tp": 0,
361
+ "fp": 0,
362
+ "fn": 723
363
+ },
364
+ "19": {
365
+ "f1_score": 0.0,
366
+ "precision": 0.0,
367
+ "recall": 0.0,
368
+ "tp": 0,
369
+ "fp": 0,
370
+ "fn": 724
371
+ },
372
+ "2": {
373
+ "f1_score": 0.0,
374
+ "precision": 0.0,
375
+ "recall": 0.0,
376
+ "tp": 0,
377
+ "fp": 0,
378
+ "fn": 725
379
+ },
380
+ "20": {
381
+ "f1_score": 0.0,
382
+ "precision": 0.0,
383
+ "recall": 0.0,
384
+ "tp": 0,
385
+ "fp": 0,
386
+ "fn": 721
387
+ },
388
+ "21": {
389
+ "f1_score": 0.0,
390
+ "precision": 0.0,
391
+ "recall": 0.0,
392
+ "tp": 0,
393
+ "fp": 0,
394
+ "fn": 725
395
+ },
396
+ "22": {
397
+ "f1_score": 0.0,
398
+ "precision": 0.0,
399
+ "recall": 0.0,
400
+ "tp": 0,
401
+ "fp": 0,
402
+ "fn": 716
403
+ },
404
+ "3": {
405
+ "f1_score": 0.0,
406
+ "precision": 0.0,
407
+ "recall": 0.0,
408
+ "tp": 0,
409
+ "fp": 0,
410
+ "fn": 724
411
+ },
412
+ "4": {
413
+ "f1_score": 0.0,
414
+ "precision": 0.0,
415
+ "recall": 0.0,
416
+ "tp": 0,
417
+ "fp": 0,
418
+ "fn": 723
419
+ },
420
+ "5": {
421
+ "f1_score": 0.0,
422
+ "precision": 0.0,
423
+ "recall": 0.0,
424
+ "tp": 0,
425
+ "fp": 0,
426
+ "fn": 723
427
+ },
428
+ "6": {
429
+ "f1_score": 0.0,
430
+ "precision": 0.0,
431
+ "recall": 0.0,
432
+ "tp": 0,
433
+ "fp": 0,
434
+ "fn": 725
435
+ },
436
+ "7": {
437
+ "f1_score": 0.0,
438
+ "precision": 0.0,
439
+ "recall": 0.0,
440
+ "tp": 0,
441
+ "fp": 0,
442
+ "fn": 722
443
+ },
444
+ "8": {
445
+ "f1_score": 0.0,
446
+ "precision": 0.0,
447
+ "recall": 0.0,
448
+ "tp": 0,
449
+ "fp": 0,
450
+ "fn": 720
451
+ },
452
+ "9": {
453
+ "f1_score": 0.0,
454
+ "precision": 0.0,
455
+ "recall": 0.0,
456
+ "tp": 0,
457
+ "fp": 0,
458
+ "fn": 724
459
+ },
460
+ "__unlabeled__": {
461
+ "f1_score": 0.005136986301369863,
462
+ "precision": 0.0028625954198473282,
463
+ "recall": 0.025,
464
+ "tp": 15,
465
+ "fp": 5225,
466
+ "fn": 585
467
+ },
468
+ "x": {
469
+ "f1_score": 0.0,
470
+ "precision": 0.0,
471
+ "recall": 0.0,
472
+ "tp": 0,
473
+ "fp": 0,
474
+ "fn": 558
475
+ },
476
+ "y": {
477
+ "f1_score": 0.0,
478
+ "precision": 0.0,
479
+ "recall": 0.0,
480
+ "tp": 0,
481
+ "fp": 0,
482
+ "fn": 174
483
+ }
484
+ },
485
+ "valid_samples": 961
486
+ },
487
+ "matching": {
488
+ "tp": 4,
489
+ "fp": 5236,
490
+ "fn": 17195
491
+ },
492
+ "iou_threshold": 0.5
493
+ },
494
+ "disease_diagnosis_classification": {
495
+ "metric": "balanced_accuracy",
496
+ "value": 0.0020100502512562816,
497
+ "count": 2610,
498
+ "metrics": {
499
+ "balanced_accuracy": 0.0020100502512562816,
500
+ "accuracy": 0.00842911877394636,
501
+ "f1_score": 0.015474120258443963,
502
+ "valid_samples": 2610
503
+ }
504
+ },
505
+ "multi_label_classification": {
506
+ "metric": "f1_score",
507
+ "value": 0.1507593198931337,
508
+ "count": 923,
509
+ "metrics": {
510
+ "f1_score": 0.1507593198931337,
511
+ "precision": 0.10261288282875378,
512
+ "recall": 0.40240158902130735,
513
+ "valid_samples": 923
514
+ }
515
+ },
516
+ "regression": {
517
+ "metric": "mean_absolute_error",
518
+ "value": 30.038603799104408,
519
+ "count": 202,
520
+ "metrics": {
521
+ "mean_absolute_error": 30.038603799104408,
522
+ "root_mean_squared_error": 46.64166890561044,
523
+ "valid_samples": 202
524
+ }
525
+ }
526
+ },
527
+ "num_rows": 4696,
528
+ "num_tasks": 4,
529
+ "predictions": "/scratch/atatc/output/medgemma-flare-2d-output/flare-medgemma-base-infer/{split_predictions_testing.jsonl}",
530
+ "split": "testing",
531
+ "model_variant": "predictions_file",
532
+ "num_predictions": 4696
533
+ },
534
+ "validation_public": {
535
+ "metrics": {
536
+ "cell_counting_mean_absolute_error": 32663.9,
537
+ "cell_counting_root_mean_squared_error": 173393.05504108287,
538
+ "f1_iou_0.5": 0.00162999185004075,
539
+ "detection_f1_iou_0.5": 0.00162999185004075,
540
+ "detection_precision_iou_0.5": 0.0009652509652509653,
541
+ "detection_recall_iou_0.5": 0.005235602094240838,
542
+ "detection_f1_iou_0.3": 0.04237978810105949,
543
+ "detection_precision_iou_0.3": 0.025096525096525095,
544
+ "detection_recall_iou_0.3": 0.13612565445026178,
545
+ "detection_average_f1": 0.012387938060309698,
546
+ "detection_average_precision": 0.007335907335907335,
547
+ "detection_average_recall": 0.03979057591623037,
548
+ "detection_iou_0.3_f1_score": 0.04237978810105949,
549
+ "detection_iou_0.3_precision": 0.025096525096525095,
550
+ "detection_iou_0.3_recall": 0.13612565445026178,
551
+ "detection_iou_0.3_tp": 26,
552
+ "detection_iou_0.3_fp": 1010,
553
+ "detection_iou_0.3_fn": 165,
554
+ "detection_iou_0.4_f1_score": 0.017929910350448247,
555
+ "detection_iou_0.4_precision": 0.010617760617760617,
556
+ "detection_iou_0.4_recall": 0.05759162303664921,
557
+ "detection_iou_0.4_tp": 11,
558
+ "detection_iou_0.4_fp": 1025,
559
+ "detection_iou_0.4_fn": 180,
560
+ "detection_iou_0.5_f1_score": 0.00162999185004075,
561
+ "detection_iou_0.5_precision": 0.0009652509652509653,
562
+ "detection_iou_0.5_recall": 0.005235602094240838,
563
+ "detection_iou_0.5_tp": 1,
564
+ "detection_iou_0.5_fp": 1035,
565
+ "detection_iou_0.5_fn": 190,
566
+ "detection_iou_0.6_f1_score": 0.0,
567
+ "detection_iou_0.6_precision": 0.0,
568
+ "detection_iou_0.6_recall": 0.0,
569
+ "detection_iou_0.6_tp": 0,
570
+ "detection_iou_0.6_fp": 1036,
571
+ "detection_iou_0.6_fn": 191,
572
+ "detection_iou_0.7_f1_score": 0.0,
573
+ "detection_iou_0.7_precision": 0.0,
574
+ "detection_iou_0.7_recall": 0.0,
575
+ "detection_iou_0.7_tp": 0,
576
+ "detection_iou_0.7_fp": 1036,
577
+ "detection_iou_0.7_fn": 191,
578
+ "balanced_accuracy": 0.041802081600207056,
579
+ "classification_accuracy": 0.06702974444909929,
580
+ "classification_f1_score": 0.06692966018987825,
581
+ "f1_score": 0.17710801631592055,
582
+ "multi_label_f1_score": 0.17710801631592055,
583
+ "multi_label_precision": 0.10996628761191016,
584
+ "multi_label_recall": 0.6910775650466372,
585
+ "green_score": 0.771889052107561,
586
+ "crimson_score": 0.69777170781893
587
+ },
588
+ "by_task": {
589
+ "cell_counting": {
590
+ "metric": "mean_absolute_error",
591
+ "value": 32663.9,
592
+ "count": 100,
593
+ "metrics": {
594
+ "mean_absolute_error": 32663.9,
595
+ "root_mean_squared_error": 173393.05504108287,
596
+ "valid_samples": 100
597
+ }
598
+ },
599
+ "detection": {
600
+ "metric": "f1_iou_0.5",
601
+ "value": 0.00162999185004075,
602
+ "count": 175,
603
+ "metrics": {
604
+ "f1_score": 0.00162999185004075,
605
+ "precision": 0.0009652509652509653,
606
+ "recall": 0.005235602094240838,
607
+ "f1_score_at_03": 0.04237978810105949,
608
+ "f1_score_at_05": 0.00162999185004075,
609
+ "average_f1": 0.012387938060309698,
610
+ "precision_at_03": 0.025096525096525095,
611
+ "recall_at_03": 0.13612565445026178,
612
+ "detailed_metrics": {
613
+ "IoU_0.3": {
614
+ "f1_score": 0.04237978810105949,
615
+ "precision": 0.025096525096525095,
616
+ "recall": 0.13612565445026178,
617
+ "tp": 26,
618
+ "fp": 1010,
619
+ "fn": 165
620
+ },
621
+ "IoU_0.4": {
622
+ "f1_score": 0.017929910350448247,
623
+ "precision": 0.010617760617760617,
624
+ "recall": 0.05759162303664921,
625
+ "tp": 11,
626
+ "fp": 1025,
627
+ "fn": 180
628
+ },
629
+ "IoU_0.5": {
630
+ "f1_score": 0.00162999185004075,
631
+ "precision": 0.0009652509652509653,
632
+ "recall": 0.005235602094240838,
633
+ "tp": 1,
634
+ "fp": 1035,
635
+ "fn": 190
636
+ },
637
+ "IoU_0.6": {
638
+ "f1_score": 0.0,
639
+ "precision": 0.0,
640
+ "recall": 0.0,
641
+ "tp": 0,
642
+ "fp": 1036,
643
+ "fn": 191
644
+ },
645
+ "IoU_0.7": {
646
+ "f1_score": 0.0,
647
+ "precision": 0.0,
648
+ "recall": 0.0,
649
+ "tp": 0,
650
+ "fp": 1036,
651
+ "fn": 191
652
+ }
653
+ },
654
+ "coco_style_metrics": {
655
+ "average_f1": 0.012387938060309698,
656
+ "average_precision": 0.007335907335907335,
657
+ "average_recall": 0.03979057591623037
658
+ },
659
+ "per_chromosome_metrics": {
660
+ "__unlabeled__": {
661
+ "f1_score": 0.04237978810105949,
662
+ "precision": 0.025096525096525095,
663
+ "recall": 0.13612565445026178,
664
+ "tp": 26,
665
+ "fp": 1010,
666
+ "fn": 165
667
+ }
668
+ },
669
+ "valid_samples": 175
670
+ },
671
+ "matching": {
672
+ "tp": 1,
673
+ "fp": 1035,
674
+ "fn": 190
675
+ },
676
+ "iou_threshold": 0.5
677
+ },
678
+ "disease_diagnosis_classification": {
679
+ "metric": "balanced_accuracy",
680
+ "value": 0.041802081600207056,
681
+ "count": 2387,
682
+ "metrics": {
683
+ "balanced_accuracy": 0.041802081600207056,
684
+ "accuracy": 0.06702974444909929,
685
+ "f1_score": 0.06692966018987825,
686
+ "valid_samples": 2387
687
+ }
688
+ },
689
+ "multi_label_classification": {
690
+ "metric": "f1_score",
691
+ "value": 0.17710801631592055,
692
+ "count": 970,
693
+ "metrics": {
694
+ "f1_score": 0.17710801631592055,
695
+ "precision": 0.10996628761191016,
696
+ "recall": 0.6910775650466372,
697
+ "valid_samples": 970
698
+ }
699
+ },
700
+ "report_generation": {
701
+ "metric": "green_score",
702
+ "value": 0.771889052107561,
703
+ "count": 1945,
704
+ "metrics": {
705
+ "green_score": 0.771889052107561,
706
+ "valid_samples": 1945,
707
+ "crimson_score": 0.69777170781893,
708
+ "crimson_valid_samples": 1944,
709
+ "crimson_failed_samples": 1,
710
+ "crimson_false_findings": 0.31121399176954734,
711
+ "crimson_missing_findings": 0.2227366255144033,
712
+ "crimson_attribute_errors": 0.09362139917695474,
713
+ "crimson_location_errors": 0.01286008230452675,
714
+ "crimson_severity_errors": 0.016975308641975308,
715
+ "crimson_descriptor_errors": 0.01131687242798354,
716
+ "crimson_measurement_errors": 0.0015432098765432098,
717
+ "crimson_certainty_errors": 0.011831275720164609,
718
+ "crimson_unspecific_errors": 0.013888888888888888,
719
+ "crimson_overinterpretation_errors": 0.010802469135802469,
720
+ "crimson_temporal_errors": 0.01440329218106996,
721
+ "crimson_weighted_false_findings": 0.13425925925925927,
722
+ "crimson_weighted_missing_findings": 0.04925411522633745,
723
+ "crimson_weighted_attribute_errors": 0.016975308641975308,
724
+ "crimson_N_G": 0.06520061728395062,
725
+ "crimson_E_penalty": 0.13425925925925927,
726
+ "crimson_correct": 0.011233588085439937,
727
+ "crimson_errors_more_than_correct": 0.12302567117381932,
728
+ "crimson_S": 0.4654436660352298
729
+ }
730
+ }
731
+ },
732
+ "num_rows": 5577,
733
+ "num_tasks": 5,
734
+ "predictions": "/scratch/atatc/output/medgemma-flare-2d-output/flare-medgemma-base-infer/{split_predictions_validation_public.jsonl}",
735
+ "split": "validation_public",
736
+ "model_variant": "predictions_file",
737
+ "num_predictions": 5577
738
+ },
739
+ "validation_hidden": {
740
+ "metrics": {
741
+ "f1_iou_0.5": 0.00019882692116512577,
742
+ "detection_f1_iou_0.5": 0.00019882692116512577,
743
+ "detection_precision_iou_0.5": 0.0005260389268805891,
744
+ "detection_recall_iou_0.5": 0.00012257906349595488,
745
+ "detection_f1_iou_0.3": 0.0013917884481558804,
746
+ "detection_precision_iou_0.3": 0.003682272488164124,
747
+ "detection_recall_iou_0.3": 0.0008580534444716843,
748
+ "detection_average_f1": 0.0003578884580972264,
749
+ "detection_average_precision": 0.0009468700683850605,
750
+ "detection_average_recall": 0.0002206423142927188,
751
+ "detection_iou_0.3_f1_score": 0.0013917884481558804,
752
+ "detection_iou_0.3_precision": 0.003682272488164124,
753
+ "detection_iou_0.3_recall": 0.0008580534444716843,
754
+ "detection_iou_0.3_tp": 7,
755
+ "detection_iou_0.3_fp": 1894,
756
+ "detection_iou_0.3_fn": 8151,
757
+ "detection_iou_0.4_f1_score": 0.00019882692116512577,
758
+ "detection_iou_0.4_precision": 0.0005260389268805891,
759
+ "detection_iou_0.4_recall": 0.00012257906349595488,
760
+ "detection_iou_0.4_tp": 1,
761
+ "detection_iou_0.4_fp": 1900,
762
+ "detection_iou_0.4_fn": 8157,
763
+ "detection_iou_0.5_f1_score": 0.00019882692116512577,
764
+ "detection_iou_0.5_precision": 0.0005260389268805891,
765
+ "detection_iou_0.5_recall": 0.00012257906349595488,
766
+ "detection_iou_0.5_tp": 1,
767
+ "detection_iou_0.5_fp": 1900,
768
+ "detection_iou_0.5_fn": 8157,
769
+ "detection_iou_0.6_f1_score": 0.0,
770
+ "detection_iou_0.6_precision": 0.0,
771
+ "detection_iou_0.6_recall": 0.0,
772
+ "detection_iou_0.6_tp": 0,
773
+ "detection_iou_0.6_fp": 1901,
774
+ "detection_iou_0.6_fn": 8158,
775
+ "detection_iou_0.7_f1_score": 0.0,
776
+ "detection_iou_0.7_precision": 0.0,
777
+ "detection_iou_0.7_recall": 0.0,
778
+ "detection_iou_0.7_tp": 0,
779
+ "detection_iou_0.7_fp": 1901,
780
+ "detection_iou_0.7_fn": 8158,
781
+ "balanced_accuracy": 0.0022377622377622378,
782
+ "classification_accuracy": 0.007104795737122558,
783
+ "classification_f1_score": 0.012515222843169818,
784
+ "f1_score": 0.134390183421731,
785
+ "multi_label_f1_score": 0.134390183421731,
786
+ "multi_label_precision": 0.09584958493883318,
787
+ "multi_label_recall": 0.35359888190076866,
788
+ "regression_mean_absolute_error": 26.376872231970015,
789
+ "regression_root_mean_squared_error": 34.00022819787254
790
+ },
791
+ "by_task": {
792
+ "detection": {
793
+ "metric": "f1_iou_0.5",
794
+ "value": 0.00019882692116512577,
795
+ "count": 256,
796
+ "metrics": {
797
+ "f1_score": 0.00019882692116512577,
798
+ "precision": 0.0005260389268805891,
799
+ "recall": 0.00012257906349595488,
800
+ "f1_score_at_03": 0.0013917884481558804,
801
+ "f1_score_at_05": 0.00019882692116512577,
802
+ "average_f1": 0.0003578884580972264,
803
+ "precision_at_03": 0.003682272488164124,
804
+ "recall_at_03": 0.0008580534444716843,
805
+ "detailed_metrics": {
806
+ "IoU_0.3": {
807
+ "f1_score": 0.0013917884481558804,
808
+ "precision": 0.003682272488164124,
809
+ "recall": 0.0008580534444716843,
810
+ "tp": 7,
811
+ "fp": 1894,
812
+ "fn": 8151
813
+ },
814
+ "IoU_0.4": {
815
+ "f1_score": 0.00019882692116512577,
816
+ "precision": 0.0005260389268805891,
817
+ "recall": 0.00012257906349595488,
818
+ "tp": 1,
819
+ "fp": 1900,
820
+ "fn": 8157
821
+ },
822
+ "IoU_0.5": {
823
+ "f1_score": 0.00019882692116512577,
824
+ "precision": 0.0005260389268805891,
825
+ "recall": 0.00012257906349595488,
826
+ "tp": 1,
827
+ "fp": 1900,
828
+ "fn": 8157
829
+ },
830
+ "IoU_0.6": {
831
+ "f1_score": 0.0,
832
+ "precision": 0.0,
833
+ "recall": 0.0,
834
+ "tp": 0,
835
+ "fp": 1901,
836
+ "fn": 8158
837
+ },
838
+ "IoU_0.7": {
839
+ "f1_score": 0.0,
840
+ "precision": 0.0,
841
+ "recall": 0.0,
842
+ "tp": 0,
843
+ "fp": 1901,
844
+ "fn": 8158
845
+ }
846
+ },
847
+ "coco_style_metrics": {
848
+ "average_f1": 0.0003578884580972264,
849
+ "average_precision": 0.0009468700683850605,
850
+ "average_recall": 0.0002206423142927188
851
+ },
852
+ "per_chromosome_metrics": {
853
+ "1": {
854
+ "f1_score": 0.0,
855
+ "precision": 0.0,
856
+ "recall": 0.0,
857
+ "tp": 0,
858
+ "fp": 0,
859
+ "fn": 352
860
+ },
861
+ "10": {
862
+ "f1_score": 0.0,
863
+ "precision": 0.0,
864
+ "recall": 0.0,
865
+ "tp": 0,
866
+ "fp": 0,
867
+ "fn": 352
868
+ },
869
+ "11": {
870
+ "f1_score": 0.0,
871
+ "precision": 0.0,
872
+ "recall": 0.0,
873
+ "tp": 0,
874
+ "fp": 0,
875
+ "fn": 353
876
+ },
877
+ "12": {
878
+ "f1_score": 0.0,
879
+ "precision": 0.0,
880
+ "recall": 0.0,
881
+ "tp": 0,
882
+ "fp": 0,
883
+ "fn": 351
884
+ },
885
+ "13": {
886
+ "f1_score": 0.0,
887
+ "precision": 0.0,
888
+ "recall": 0.0,
889
+ "tp": 0,
890
+ "fp": 0,
891
+ "fn": 349
892
+ },
893
+ "14": {
894
+ "f1_score": 0.0,
895
+ "precision": 0.0,
896
+ "recall": 0.0,
897
+ "tp": 0,
898
+ "fp": 0,
899
+ "fn": 340
900
+ },
901
+ "15": {
902
+ "f1_score": 0.0,
903
+ "precision": 0.0,
904
+ "recall": 0.0,
905
+ "tp": 0,
906
+ "fp": 0,
907
+ "fn": 351
908
+ },
909
+ "16": {
910
+ "f1_score": 0.0,
911
+ "precision": 0.0,
912
+ "recall": 0.0,
913
+ "tp": 0,
914
+ "fp": 0,
915
+ "fn": 352
916
+ },
917
+ "17": {
918
+ "f1_score": 0.0,
919
+ "precision": 0.0,
920
+ "recall": 0.0,
921
+ "tp": 0,
922
+ "fp": 0,
923
+ "fn": 350
924
+ },
925
+ "18": {
926
+ "f1_score": 0.0,
927
+ "precision": 0.0,
928
+ "recall": 0.0,
929
+ "tp": 0,
930
+ "fp": 0,
931
+ "fn": 353
932
+ },
933
+ "19": {
934
+ "f1_score": 0.0,
935
+ "precision": 0.0,
936
+ "recall": 0.0,
937
+ "tp": 0,
938
+ "fp": 0,
939
+ "fn": 352
940
+ },
941
+ "2": {
942
+ "f1_score": 0.0,
943
+ "precision": 0.0,
944
+ "recall": 0.0,
945
+ "tp": 0,
946
+ "fp": 0,
947
+ "fn": 352
948
+ },
949
+ "20": {
950
+ "f1_score": 0.0,
951
+ "precision": 0.0,
952
+ "recall": 0.0,
953
+ "tp": 0,
954
+ "fp": 0,
955
+ "fn": 354
956
+ },
957
+ "21": {
958
+ "f1_score": 0.0,
959
+ "precision": 0.0,
960
+ "recall": 0.0,
961
+ "tp": 0,
962
+ "fp": 0,
963
+ "fn": 357
964
+ },
965
+ "22": {
966
+ "f1_score": 0.0,
967
+ "precision": 0.0,
968
+ "recall": 0.0,
969
+ "tp": 0,
970
+ "fp": 0,
971
+ "fn": 350
972
+ },
973
+ "3": {
974
+ "f1_score": 0.0,
975
+ "precision": 0.0,
976
+ "recall": 0.0,
977
+ "tp": 0,
978
+ "fp": 0,
979
+ "fn": 350
980
+ },
981
+ "4": {
982
+ "f1_score": 0.0,
983
+ "precision": 0.0,
984
+ "recall": 0.0,
985
+ "tp": 0,
986
+ "fp": 0,
987
+ "fn": 351
988
+ },
989
+ "5": {
990
+ "f1_score": 0.0,
991
+ "precision": 0.0,
992
+ "recall": 0.0,
993
+ "tp": 0,
994
+ "fp": 0,
995
+ "fn": 350
996
+ },
997
+ "6": {
998
+ "f1_score": 0.0,
999
+ "precision": 0.0,
1000
+ "recall": 0.0,
1001
+ "tp": 0,
1002
+ "fp": 0,
1003
+ "fn": 353
1004
+ },
1005
+ "7": {
1006
+ "f1_score": 0.0,
1007
+ "precision": 0.0,
1008
+ "recall": 0.0,
1009
+ "tp": 0,
1010
+ "fp": 0,
1011
+ "fn": 353
1012
+ },
1013
+ "8": {
1014
+ "f1_score": 0.0,
1015
+ "precision": 0.0,
1016
+ "recall": 0.0,
1017
+ "tp": 0,
1018
+ "fp": 0,
1019
+ "fn": 351
1020
+ },
1021
+ "9": {
1022
+ "f1_score": 0.0,
1023
+ "precision": 0.0,
1024
+ "recall": 0.0,
1025
+ "tp": 0,
1026
+ "fp": 0,
1027
+ "fn": 349
1028
+ },
1029
+ "__unlabeled__": {
1030
+ "f1_score": 0.0,
1031
+ "precision": 0.0,
1032
+ "recall": 0.0,
1033
+ "tp": 0,
1034
+ "fp": 1901,
1035
+ "fn": 80
1036
+ },
1037
+ "x": {
1038
+ "f1_score": 0.0,
1039
+ "precision": 0.0,
1040
+ "recall": 0.0,
1041
+ "tp": 0,
1042
+ "fp": 0,
1043
+ "fn": 277
1044
+ },
1045
+ "y": {
1046
+ "f1_score": 0.0,
1047
+ "precision": 0.0,
1048
+ "recall": 0.0,
1049
+ "tp": 0,
1050
+ "fp": 0,
1051
+ "fn": 76
1052
+ }
1053
+ },
1054
+ "valid_samples": 256
1055
+ },
1056
+ "matching": {
1057
+ "tp": 1,
1058
+ "fp": 1900,
1059
+ "fn": 8157
1060
+ },
1061
+ "iou_threshold": 0.5
1062
+ },
1063
+ "disease_diagnosis_classification": {
1064
+ "metric": "balanced_accuracy",
1065
+ "value": 0.0022377622377622378,
1066
+ "count": 1126,
1067
+ "metrics": {
1068
+ "balanced_accuracy": 0.0022377622377622378,
1069
+ "accuracy": 0.007104795737122558,
1070
+ "f1_score": 0.012515222843169818,
1071
+ "valid_samples": 1126
1072
+ }
1073
+ },
1074
+ "multi_label_classification": {
1075
+ "metric": "f1_score",
1076
+ "value": 0.134390183421731,
1077
+ "count": 477,
1078
+ "metrics": {
1079
+ "f1_score": 0.134390183421731,
1080
+ "precision": 0.09584958493883318,
1081
+ "recall": 0.35359888190076866,
1082
+ "valid_samples": 477
1083
+ }
1084
+ },
1085
+ "regression": {
1086
+ "metric": "mean_absolute_error",
1087
+ "value": 26.376872231970015,
1088
+ "count": 100,
1089
+ "metrics": {
1090
+ "mean_absolute_error": 26.376872231970015,
1091
+ "root_mean_squared_error": 34.00022819787254,
1092
+ "valid_samples": 100
1093
+ }
1094
+ }
1095
+ },
1096
+ "num_rows": 1959,
1097
+ "num_tasks": 4,
1098
+ "predictions": "/scratch/atatc/output/medgemma-flare-2d-output/flare-medgemma-base-infer/{split_predictions_validation_hidden.jsonl}",
1099
+ "split": "validation_hidden",
1100
+ "model_variant": "predictions_file",
1101
+ "num_predictions": 3918
1102
+ }
1103
+ },
1104
+ "num_rows": 12232,
1105
+ "num_splits": 3,
1106
+ "num_tasks": 6,
1107
+ "splits": [
1108
+ "testing",
1109
+ "validation_public",
1110
+ "validation_hidden"
1111
+ ],
1112
+ "model_variant": "predictions_file"
1113
+ }
flare-medgemma-base-eval/scores.json ADDED
@@ -0,0 +1,214 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "mean": {
3
+ "balanced_accuracy": 0.015349964696408524,
4
+ "cell_counting_mean_absolute_error": 32663.9,
5
+ "cell_counting_root_mean_squared_error": 173393.05504108287,
6
+ "classification_accuracy": 0.0275212196533894,
7
+ "classification_f1_score": 0.03163966776383068,
8
+ "crimson_score": 0.69777170781893,
9
+ "detection_average_f1": 0.004551652647125287,
10
+ "detection_average_precision": 0.00340978076326286,
11
+ "detection_average_recall": 0.013534758655479323,
12
+ "detection_f1_iou_0.3": 0.015570961364704957,
13
+ "detection_f1_iou_0.5": 0.0007284469659534537,
14
+ "detection_iou_0.3_f1_score": 0.015570961364704957,
15
+ "detection_iou_0.3_fn": 8494.0,
16
+ "detection_iou_0.3_fp": 2703.6666666666665,
17
+ "detection_iou_0.3_precision": 0.011692169169451113,
18
+ "detection_iou_0.3_recall": 0.04630080803305465,
19
+ "detection_iou_0.3_tp": 22.0,
20
+ "detection_iou_0.4_f1_score": 0.00639943455052561,
21
+ "detection_iou_0.4_fn": 8508.0,
22
+ "detection_iou_0.4_fp": 2717.6666666666665,
23
+ "detection_iou_0.4_precision": 0.004477958626839689,
24
+ "detection_iou_0.4_recall": 0.019470639027858148,
25
+ "detection_iou_0.4_tp": 8.0,
26
+ "detection_iou_0.5_f1_score": 0.0007284469659534537,
27
+ "detection_iou_0.5_fn": 8514.0,
28
+ "detection_iou_0.5_fp": 2723.6666666666665,
29
+ "detection_iou_0.5_precision": 0.0007515495569191695,
30
+ "detection_iou_0.5_recall": 0.0018635842729599608,
31
+ "detection_iou_0.5_tp": 2.0,
32
+ "detection_iou_0.6_f1_score": 2.971017722120712e-05,
33
+ "detection_iou_0.6_fn": 8515.666666666666,
34
+ "detection_iou_0.6_fp": 2725.3333333333335,
35
+ "detection_iou_0.6_precision": 6.361323155216285e-05,
36
+ "detection_iou_0.6_recall": 1.938097176192414e-05,
37
+ "detection_iou_0.6_tp": 0.3333333333333333,
38
+ "detection_iou_0.7_f1_score": 2.971017722120712e-05,
39
+ "detection_iou_0.7_fn": 8515.666666666666,
40
+ "detection_iou_0.7_fp": 2725.3333333333335,
41
+ "detection_iou_0.7_precision": 6.361323155216285e-05,
42
+ "detection_iou_0.7_recall": 1.938097176192414e-05,
43
+ "detection_iou_0.7_tp": 0.3333333333333333,
44
+ "detection_precision_iou_0.3": 0.011692169169451113,
45
+ "detection_precision_iou_0.5": 0.0007515495569191695,
46
+ "detection_recall_iou_0.3": 0.04630080803305465,
47
+ "detection_recall_iou_0.5": 0.0018635842729599608,
48
+ "f1_iou_0.5": 0.0007284469659534537,
49
+ "f1_score": 0.15408583987692842,
50
+ "green_score": 0.771889052107561,
51
+ "multi_label_f1_score": 0.15408583987692842,
52
+ "multi_label_precision": 0.10280958512649903,
53
+ "multi_label_recall": 0.4823593453229044,
54
+ "regression_mean_absolute_error": 28.20773801553721,
55
+ "regression_root_mean_squared_error": 40.32094855174149
56
+ },
57
+ "per_split": {
58
+ "testing": {
59
+ "f1_iou_0.5": 0.0003565221266544854,
60
+ "detection_f1_iou_0.5": 0.0003565221266544854,
61
+ "detection_precision_iou_0.5": 0.0007633587786259542,
62
+ "detection_recall_iou_0.5": 0.0002325716611430897,
63
+ "detection_f1_iou_0.3": 0.0029413075448995055,
64
+ "detection_precision_iou_0.3": 0.006297709923664122,
65
+ "detection_recall_iou_0.3": 0.0019187162044304901,
66
+ "detection_average_f1": 0.000909131422968938,
67
+ "detection_average_precision": 0.0019465648854961833,
68
+ "detection_average_recall": 0.0005930577359148788,
69
+ "detection_iou_0.3_f1_score": 0.0029413075448995055,
70
+ "detection_iou_0.3_precision": 0.006297709923664122,
71
+ "detection_iou_0.3_recall": 0.0019187162044304901,
72
+ "detection_iou_0.3_tp": 33,
73
+ "detection_iou_0.3_fp": 5207,
74
+ "detection_iou_0.3_fn": 17166,
75
+ "detection_iou_0.4_f1_score": 0.0010695663799634564,
76
+ "detection_iou_0.4_precision": 0.0022900763358778627,
77
+ "detection_iou_0.4_recall": 0.0006977149834292691,
78
+ "detection_iou_0.4_tp": 12,
79
+ "detection_iou_0.4_fp": 5228,
80
+ "detection_iou_0.4_fn": 17187,
81
+ "detection_iou_0.5_f1_score": 0.0003565221266544854,
82
+ "detection_iou_0.5_precision": 0.0007633587786259542,
83
+ "detection_iou_0.5_recall": 0.0002325716611430897,
84
+ "detection_iou_0.5_tp": 4,
85
+ "detection_iou_0.5_fp": 5236,
86
+ "detection_iou_0.5_fn": 17195,
87
+ "detection_iou_0.6_f1_score": 8.913053166362135e-05,
88
+ "detection_iou_0.6_precision": 0.00019083969465648855,
89
+ "detection_iou_0.6_recall": 5.8142915285772426e-05,
90
+ "detection_iou_0.6_tp": 1,
91
+ "detection_iou_0.6_fp": 5239,
92
+ "detection_iou_0.6_fn": 17198,
93
+ "detection_iou_0.7_f1_score": 8.913053166362135e-05,
94
+ "detection_iou_0.7_precision": 0.00019083969465648855,
95
+ "detection_iou_0.7_recall": 5.8142915285772426e-05,
96
+ "detection_iou_0.7_tp": 1,
97
+ "detection_iou_0.7_fp": 5239,
98
+ "detection_iou_0.7_fn": 17198,
99
+ "balanced_accuracy": 0.0020100502512562816,
100
+ "classification_accuracy": 0.00842911877394636,
101
+ "classification_f1_score": 0.015474120258443963,
102
+ "f1_score": 0.1507593198931337,
103
+ "multi_label_f1_score": 0.1507593198931337,
104
+ "multi_label_precision": 0.10261288282875378,
105
+ "multi_label_recall": 0.40240158902130735,
106
+ "regression_mean_absolute_error": 30.038603799104408,
107
+ "regression_root_mean_squared_error": 46.64166890561044
108
+ },
109
+ "validation_public": {
110
+ "cell_counting_mean_absolute_error": 32663.9,
111
+ "cell_counting_root_mean_squared_error": 173393.05504108287,
112
+ "f1_iou_0.5": 0.00162999185004075,
113
+ "detection_f1_iou_0.5": 0.00162999185004075,
114
+ "detection_precision_iou_0.5": 0.0009652509652509653,
115
+ "detection_recall_iou_0.5": 0.005235602094240838,
116
+ "detection_f1_iou_0.3": 0.04237978810105949,
117
+ "detection_precision_iou_0.3": 0.025096525096525095,
118
+ "detection_recall_iou_0.3": 0.13612565445026178,
119
+ "detection_average_f1": 0.012387938060309698,
120
+ "detection_average_precision": 0.007335907335907335,
121
+ "detection_average_recall": 0.03979057591623037,
122
+ "detection_iou_0.3_f1_score": 0.04237978810105949,
123
+ "detection_iou_0.3_precision": 0.025096525096525095,
124
+ "detection_iou_0.3_recall": 0.13612565445026178,
125
+ "detection_iou_0.3_tp": 26,
126
+ "detection_iou_0.3_fp": 1010,
127
+ "detection_iou_0.3_fn": 165,
128
+ "detection_iou_0.4_f1_score": 0.017929910350448247,
129
+ "detection_iou_0.4_precision": 0.010617760617760617,
130
+ "detection_iou_0.4_recall": 0.05759162303664921,
131
+ "detection_iou_0.4_tp": 11,
132
+ "detection_iou_0.4_fp": 1025,
133
+ "detection_iou_0.4_fn": 180,
134
+ "detection_iou_0.5_f1_score": 0.00162999185004075,
135
+ "detection_iou_0.5_precision": 0.0009652509652509653,
136
+ "detection_iou_0.5_recall": 0.005235602094240838,
137
+ "detection_iou_0.5_tp": 1,
138
+ "detection_iou_0.5_fp": 1035,
139
+ "detection_iou_0.5_fn": 190,
140
+ "detection_iou_0.6_f1_score": 0.0,
141
+ "detection_iou_0.6_precision": 0.0,
142
+ "detection_iou_0.6_recall": 0.0,
143
+ "detection_iou_0.6_tp": 0,
144
+ "detection_iou_0.6_fp": 1036,
145
+ "detection_iou_0.6_fn": 191,
146
+ "detection_iou_0.7_f1_score": 0.0,
147
+ "detection_iou_0.7_precision": 0.0,
148
+ "detection_iou_0.7_recall": 0.0,
149
+ "detection_iou_0.7_tp": 0,
150
+ "detection_iou_0.7_fp": 1036,
151
+ "detection_iou_0.7_fn": 191,
152
+ "balanced_accuracy": 0.041802081600207056,
153
+ "classification_accuracy": 0.06702974444909929,
154
+ "classification_f1_score": 0.06692966018987825,
155
+ "f1_score": 0.17710801631592055,
156
+ "multi_label_f1_score": 0.17710801631592055,
157
+ "multi_label_precision": 0.10996628761191016,
158
+ "multi_label_recall": 0.6910775650466372,
159
+ "green_score": 0.771889052107561,
160
+ "crimson_score": 0.69777170781893
161
+ },
162
+ "validation_hidden": {
163
+ "f1_iou_0.5": 0.00019882692116512577,
164
+ "detection_f1_iou_0.5": 0.00019882692116512577,
165
+ "detection_precision_iou_0.5": 0.0005260389268805891,
166
+ "detection_recall_iou_0.5": 0.00012257906349595488,
167
+ "detection_f1_iou_0.3": 0.0013917884481558804,
168
+ "detection_precision_iou_0.3": 0.003682272488164124,
169
+ "detection_recall_iou_0.3": 0.0008580534444716843,
170
+ "detection_average_f1": 0.0003578884580972264,
171
+ "detection_average_precision": 0.0009468700683850605,
172
+ "detection_average_recall": 0.0002206423142927188,
173
+ "detection_iou_0.3_f1_score": 0.0013917884481558804,
174
+ "detection_iou_0.3_precision": 0.003682272488164124,
175
+ "detection_iou_0.3_recall": 0.0008580534444716843,
176
+ "detection_iou_0.3_tp": 7,
177
+ "detection_iou_0.3_fp": 1894,
178
+ "detection_iou_0.3_fn": 8151,
179
+ "detection_iou_0.4_f1_score": 0.00019882692116512577,
180
+ "detection_iou_0.4_precision": 0.0005260389268805891,
181
+ "detection_iou_0.4_recall": 0.00012257906349595488,
182
+ "detection_iou_0.4_tp": 1,
183
+ "detection_iou_0.4_fp": 1900,
184
+ "detection_iou_0.4_fn": 8157,
185
+ "detection_iou_0.5_f1_score": 0.00019882692116512577,
186
+ "detection_iou_0.5_precision": 0.0005260389268805891,
187
+ "detection_iou_0.5_recall": 0.00012257906349595488,
188
+ "detection_iou_0.5_tp": 1,
189
+ "detection_iou_0.5_fp": 1900,
190
+ "detection_iou_0.5_fn": 8157,
191
+ "detection_iou_0.6_f1_score": 0.0,
192
+ "detection_iou_0.6_precision": 0.0,
193
+ "detection_iou_0.6_recall": 0.0,
194
+ "detection_iou_0.6_tp": 0,
195
+ "detection_iou_0.6_fp": 1901,
196
+ "detection_iou_0.6_fn": 8158,
197
+ "detection_iou_0.7_f1_score": 0.0,
198
+ "detection_iou_0.7_precision": 0.0,
199
+ "detection_iou_0.7_recall": 0.0,
200
+ "detection_iou_0.7_tp": 0,
201
+ "detection_iou_0.7_fp": 1901,
202
+ "detection_iou_0.7_fn": 8158,
203
+ "balanced_accuracy": 0.0022377622377622378,
204
+ "classification_accuracy": 0.007104795737122558,
205
+ "classification_f1_score": 0.012515222843169818,
206
+ "f1_score": 0.134390183421731,
207
+ "multi_label_f1_score": 0.134390183421731,
208
+ "multi_label_precision": 0.09584958493883318,
209
+ "multi_label_recall": 0.35359888190076866,
210
+ "regression_mean_absolute_error": 26.376872231970015,
211
+ "regression_root_mean_squared_error": 34.00022819787254
212
+ }
213
+ }
214
+ }
flare-medgemma-base-infer/inference_details.json ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "splits": [
3
+ "testing",
4
+ "validation_public",
5
+ "validation_hidden"
6
+ ],
7
+ "tasks": [
8
+ "disease_diagnosis_classification",
9
+ "multi_label_classification",
10
+ "detection",
11
+ "cell_counting",
12
+ "regression",
13
+ "report_generation"
14
+ ],
15
+ "model_variant": "base",
16
+ "split_results": {
17
+ "testing": {
18
+ "split": "testing",
19
+ "predictions": "/scratch/atatc/output/medgemma-flare-2d-output/flare-medgemma-base-infer/{split_predictions_testing.jsonl}",
20
+ "num_predictions": 4696,
21
+ "num_rows": 4696,
22
+ "model_variant": "base"
23
+ },
24
+ "validation_public": {
25
+ "split": "validation_public",
26
+ "predictions": "/scratch/atatc/output/medgemma-flare-2d-output/flare-medgemma-base-infer/{split_predictions_validation_public.jsonl}",
27
+ "num_predictions": 5577,
28
+ "num_rows": 5577,
29
+ "model_variant": "base"
30
+ },
31
+ "validation_hidden": {
32
+ "split": "validation_hidden",
33
+ "predictions": "/scratch/atatc/output/medgemma-flare-2d-output/flare-medgemma-base-infer/{split_predictions_validation_hidden.jsonl}",
34
+ "num_predictions": 3918,
35
+ "num_rows": 3918,
36
+ "model_variant": "base"
37
+ }
38
+ },
39
+ "num_predictions": 14191
40
+ }
flare-medgemma-base-infer/testing_predictions.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
flare-medgemma-base-infer/validation_hidden_predictions.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
flare-medgemma-base-infer/validation_public_predictions.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
flare-medgemma-eval/details.json ADDED
@@ -0,0 +1,1113 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metrics": {
3
+ "balanced_accuracy": 0.6973681283063976,
4
+ "cell_counting_mean_absolute_error": 275.64,
5
+ "cell_counting_root_mean_squared_error": 746.0054289346693,
6
+ "classification_accuracy": 0.7545310656900996,
7
+ "classification_f1_score": 0.7486724762961748,
8
+ "crimson_score": 0.8560994341563786,
9
+ "detection_average_f1": 0.11028574193986596,
10
+ "detection_average_precision": 0.16379329963340505,
11
+ "detection_average_recall": 0.09468440474704874,
12
+ "detection_f1_iou_0.3": 0.22629028834032594,
13
+ "detection_f1_iou_0.5": 0.09221872508355598,
14
+ "detection_iou_0.3_f1_score": 0.22629028834032594,
15
+ "detection_iou_0.3_fn": 7786.666666666667,
16
+ "detection_iou_0.3_fp": 1517.3333333333333,
17
+ "detection_iou_0.3_precision": 0.35253317564260894,
18
+ "detection_iou_0.3_recall": 0.19038678870421302,
19
+ "detection_iou_0.3_tp": 729.3333333333334,
20
+ "detection_iou_0.4_f1_score": 0.1537013168992954,
21
+ "detection_iou_0.4_fn": 8090.333333333333,
22
+ "detection_iou_0.4_fp": 1821.0,
23
+ "detection_iou_0.4_precision": 0.22559317675050247,
24
+ "detection_iou_0.4_recall": 0.13258690529237713,
25
+ "detection_iou_0.4_tp": 425.6666666666667,
26
+ "detection_iou_0.5_f1_score": 0.09221872508355598,
27
+ "detection_iou_0.5_fn": 8279.0,
28
+ "detection_iou_0.5_fp": 2009.6666666666667,
29
+ "detection_iou_0.5_precision": 0.13114137416927418,
30
+ "detection_iou_0.5_recall": 0.08053482746596542,
31
+ "detection_iou_0.5_tp": 237.0,
32
+ "detection_iou_0.6_f1_score": 0.05309723809769417,
33
+ "detection_iou_0.6_fn": 8392.333333333334,
34
+ "detection_iou_0.6_fp": 2123.0,
35
+ "detection_iou_0.6_precision": 0.07378962181964394,
36
+ "detection_iou_0.6_recall": 0.04679075692400229,
37
+ "detection_iou_0.6_tp": 123.66666666666667,
38
+ "detection_iou_0.7_f1_score": 0.026121141278458215,
39
+ "detection_iou_0.7_fn": 8460.333333333334,
40
+ "detection_iou_0.7_fp": 2191.0,
41
+ "detection_iou_0.7_precision": 0.03590914978499583,
42
+ "detection_iou_0.7_recall": 0.023122745348685792,
43
+ "detection_iou_0.7_tp": 55.666666666666664,
44
+ "detection_precision_iou_0.3": 0.35253317564260894,
45
+ "detection_precision_iou_0.5": 0.13114137416927418,
46
+ "detection_recall_iou_0.3": 0.19038678870421302,
47
+ "detection_recall_iou_0.5": 0.08053482746596542,
48
+ "f1_iou_0.5": 0.09221872508355598,
49
+ "f1_score": 0.4987596735442303,
50
+ "green_score": 0.7304914922267108,
51
+ "multi_label_f1_score": 0.4987596735442303,
52
+ "multi_label_precision": 0.5195215893979218,
53
+ "multi_label_recall": 0.5108684175072767,
54
+ "regression_mean_absolute_error": 18.00935926293335,
55
+ "regression_root_mean_squared_error": 28.093081829946215
56
+ },
57
+ "mean_metrics": {
58
+ "balanced_accuracy": 0.6973681283063976,
59
+ "cell_counting_mean_absolute_error": 275.64,
60
+ "cell_counting_root_mean_squared_error": 746.0054289346693,
61
+ "classification_accuracy": 0.7545310656900996,
62
+ "classification_f1_score": 0.7486724762961748,
63
+ "crimson_score": 0.8560994341563786,
64
+ "detection_average_f1": 0.11028574193986596,
65
+ "detection_average_precision": 0.16379329963340505,
66
+ "detection_average_recall": 0.09468440474704874,
67
+ "detection_f1_iou_0.3": 0.22629028834032594,
68
+ "detection_f1_iou_0.5": 0.09221872508355598,
69
+ "detection_iou_0.3_f1_score": 0.22629028834032594,
70
+ "detection_iou_0.3_fn": 7786.666666666667,
71
+ "detection_iou_0.3_fp": 1517.3333333333333,
72
+ "detection_iou_0.3_precision": 0.35253317564260894,
73
+ "detection_iou_0.3_recall": 0.19038678870421302,
74
+ "detection_iou_0.3_tp": 729.3333333333334,
75
+ "detection_iou_0.4_f1_score": 0.1537013168992954,
76
+ "detection_iou_0.4_fn": 8090.333333333333,
77
+ "detection_iou_0.4_fp": 1821.0,
78
+ "detection_iou_0.4_precision": 0.22559317675050247,
79
+ "detection_iou_0.4_recall": 0.13258690529237713,
80
+ "detection_iou_0.4_tp": 425.6666666666667,
81
+ "detection_iou_0.5_f1_score": 0.09221872508355598,
82
+ "detection_iou_0.5_fn": 8279.0,
83
+ "detection_iou_0.5_fp": 2009.6666666666667,
84
+ "detection_iou_0.5_precision": 0.13114137416927418,
85
+ "detection_iou_0.5_recall": 0.08053482746596542,
86
+ "detection_iou_0.5_tp": 237.0,
87
+ "detection_iou_0.6_f1_score": 0.05309723809769417,
88
+ "detection_iou_0.6_fn": 8392.333333333334,
89
+ "detection_iou_0.6_fp": 2123.0,
90
+ "detection_iou_0.6_precision": 0.07378962181964394,
91
+ "detection_iou_0.6_recall": 0.04679075692400229,
92
+ "detection_iou_0.6_tp": 123.66666666666667,
93
+ "detection_iou_0.7_f1_score": 0.026121141278458215,
94
+ "detection_iou_0.7_fn": 8460.333333333334,
95
+ "detection_iou_0.7_fp": 2191.0,
96
+ "detection_iou_0.7_precision": 0.03590914978499583,
97
+ "detection_iou_0.7_recall": 0.023122745348685792,
98
+ "detection_iou_0.7_tp": 55.666666666666664,
99
+ "detection_precision_iou_0.3": 0.35253317564260894,
100
+ "detection_precision_iou_0.5": 0.13114137416927418,
101
+ "detection_recall_iou_0.3": 0.19038678870421302,
102
+ "detection_recall_iou_0.5": 0.08053482746596542,
103
+ "f1_iou_0.5": 0.09221872508355598,
104
+ "f1_score": 0.4987596735442303,
105
+ "green_score": 0.7304914922267108,
106
+ "multi_label_f1_score": 0.4987596735442303,
107
+ "multi_label_precision": 0.5195215893979218,
108
+ "multi_label_recall": 0.5108684175072767,
109
+ "regression_mean_absolute_error": 18.00935926293335,
110
+ "regression_root_mean_squared_error": 28.093081829946215
111
+ },
112
+ "by_task": {
113
+ "cell_counting": {
114
+ "metric": "mean_absolute_error",
115
+ "value": 275.64,
116
+ "count": 100,
117
+ "per_split": {
118
+ "validation_public": 275.64
119
+ }
120
+ },
121
+ "detection": {
122
+ "metric": "f1_iou_0.5",
123
+ "value": 0.09221872508355598,
124
+ "count": 1392,
125
+ "per_split": {
126
+ "testing": 0.04956549726424204,
127
+ "validation_public": 0.20054200542005424,
128
+ "validation_hidden": 0.02654867256637168
129
+ }
130
+ },
131
+ "disease_diagnosis_classification": {
132
+ "metric": "balanced_accuracy",
133
+ "value": 0.6973681283063976,
134
+ "count": 6123,
135
+ "per_split": {
136
+ "testing": 0.7543050809989986,
137
+ "validation_public": 0.5546660740446607,
138
+ "validation_hidden": 0.7831332298755335
139
+ }
140
+ },
141
+ "multi_label_classification": {
142
+ "metric": "f1_score",
143
+ "value": 0.4987596735442303,
144
+ "count": 2370,
145
+ "per_split": {
146
+ "testing": 0.4437961099932931,
147
+ "validation_public": 0.6075492977812565,
148
+ "validation_hidden": 0.44493361285814115
149
+ }
150
+ },
151
+ "regression": {
152
+ "metric": "mean_absolute_error",
153
+ "value": 18.00935926293335,
154
+ "count": 302,
155
+ "per_split": {
156
+ "testing": 20.515590607695504,
157
+ "validation_hidden": 15.5031279181712
158
+ }
159
+ },
160
+ "report_generation": {
161
+ "metric": "green_score",
162
+ "value": 0.7304914922267108,
163
+ "count": 1945,
164
+ "per_split": {
165
+ "validation_public": 0.7304914922267108
166
+ }
167
+ }
168
+ },
169
+ "split_results": {
170
+ "testing": {
171
+ "metrics": {
172
+ "f1_iou_0.5": 0.04956549726424204,
173
+ "detection_f1_iou_0.5": 0.04956549726424204,
174
+ "detection_precision_iou_0.5": 0.11846153846153847,
175
+ "detection_recall_iou_0.5": 0.03133903133903134,
176
+ "detection_f1_iou_0.3": 0.14409857924502276,
177
+ "detection_precision_iou_0.3": 0.3443956043956044,
178
+ "detection_recall_iou_0.3": 0.0911099482528054,
179
+ "detection_average_f1": 0.06310175180468068,
180
+ "detection_average_precision": 0.1508131868131868,
181
+ "detection_average_recall": 0.03989766846909704,
182
+ "detection_iou_0.3_f1_score": 0.14409857924502276,
183
+ "detection_iou_0.3_precision": 0.3443956043956044,
184
+ "detection_iou_0.3_recall": 0.0911099482528054,
185
+ "detection_iou_0.3_tp": 1567,
186
+ "detection_iou_0.3_fp": 2983,
187
+ "detection_iou_0.3_fn": 15632,
188
+ "detection_iou_0.4_f1_score": 0.08625683939491473,
189
+ "detection_iou_0.4_precision": 0.20615384615384616,
190
+ "detection_iou_0.4_recall": 0.05453805453805454,
191
+ "detection_iou_0.4_tp": 938,
192
+ "detection_iou_0.4_fp": 3612,
193
+ "detection_iou_0.4_fn": 16261,
194
+ "detection_iou_0.5_f1_score": 0.04956549726424204,
195
+ "detection_iou_0.5_precision": 0.11846153846153847,
196
+ "detection_iou_0.5_recall": 0.03133903133903134,
197
+ "detection_iou_0.5_tp": 539,
198
+ "detection_iou_0.5_fp": 4011,
199
+ "detection_iou_0.5_fn": 16660,
200
+ "detection_iou_0.6_f1_score": 0.02510460251046025,
201
+ "detection_iou_0.6_precision": 0.06,
202
+ "detection_iou_0.6_recall": 0.015873015873015872,
203
+ "detection_iou_0.6_tp": 273,
204
+ "detection_iou_0.6_fp": 4277,
205
+ "detection_iou_0.6_fn": 16926,
206
+ "detection_iou_0.7_f1_score": 0.01048324060876362,
207
+ "detection_iou_0.7_precision": 0.025054945054945054,
208
+ "detection_iou_0.7_recall": 0.006628292342578057,
209
+ "detection_iou_0.7_tp": 114,
210
+ "detection_iou_0.7_fp": 4436,
211
+ "detection_iou_0.7_fn": 17085,
212
+ "balanced_accuracy": 0.7543050809989986,
213
+ "classification_accuracy": 0.7513409961685824,
214
+ "classification_f1_score": 0.7497215228330598,
215
+ "f1_score": 0.4437961099932931,
216
+ "multi_label_f1_score": 0.4437961099932931,
217
+ "multi_label_precision": 0.45774647887323944,
218
+ "multi_label_recall": 0.45521849042975804,
219
+ "regression_mean_absolute_error": 20.515590607695504,
220
+ "regression_root_mean_squared_error": 35.72048318261565
221
+ },
222
+ "by_task": {
223
+ "detection": {
224
+ "metric": "f1_iou_0.5",
225
+ "value": 0.04956549726424204,
226
+ "count": 961,
227
+ "metrics": {
228
+ "f1_score": 0.04956549726424204,
229
+ "precision": 0.11846153846153847,
230
+ "recall": 0.03133903133903134,
231
+ "f1_score_at_03": 0.14409857924502276,
232
+ "f1_score_at_05": 0.04956549726424204,
233
+ "average_f1": 0.06310175180468068,
234
+ "precision_at_03": 0.3443956043956044,
235
+ "recall_at_03": 0.0911099482528054,
236
+ "detailed_metrics": {
237
+ "IoU_0.3": {
238
+ "f1_score": 0.14409857924502276,
239
+ "precision": 0.3443956043956044,
240
+ "recall": 0.0911099482528054,
241
+ "tp": 1567,
242
+ "fp": 2983,
243
+ "fn": 15632
244
+ },
245
+ "IoU_0.4": {
246
+ "f1_score": 0.08625683939491473,
247
+ "precision": 0.20615384615384616,
248
+ "recall": 0.05453805453805454,
249
+ "tp": 938,
250
+ "fp": 3612,
251
+ "fn": 16261
252
+ },
253
+ "IoU_0.5": {
254
+ "f1_score": 0.04956549726424204,
255
+ "precision": 0.11846153846153847,
256
+ "recall": 0.03133903133903134,
257
+ "tp": 539,
258
+ "fp": 4011,
259
+ "fn": 16660
260
+ },
261
+ "IoU_0.6": {
262
+ "f1_score": 0.02510460251046025,
263
+ "precision": 0.06,
264
+ "recall": 0.015873015873015872,
265
+ "tp": 273,
266
+ "fp": 4277,
267
+ "fn": 16926
268
+ },
269
+ "IoU_0.7": {
270
+ "f1_score": 0.01048324060876362,
271
+ "precision": 0.025054945054945054,
272
+ "recall": 0.006628292342578057,
273
+ "tp": 114,
274
+ "fp": 4436,
275
+ "fn": 17085
276
+ }
277
+ },
278
+ "coco_style_metrics": {
279
+ "average_f1": 0.06310175180468068,
280
+ "average_precision": 0.1508131868131868,
281
+ "average_recall": 0.03989766846909704
282
+ },
283
+ "per_chromosome_metrics": {
284
+ "1": {
285
+ "f1_score": 0.0,
286
+ "precision": 0.0,
287
+ "recall": 0.0,
288
+ "tp": 0,
289
+ "fp": 0,
290
+ "fn": 725
291
+ },
292
+ "10": {
293
+ "f1_score": 0.0,
294
+ "precision": 0.0,
295
+ "recall": 0.0,
296
+ "tp": 0,
297
+ "fp": 0,
298
+ "fn": 722
299
+ },
300
+ "11": {
301
+ "f1_score": 0.0,
302
+ "precision": 0.0,
303
+ "recall": 0.0,
304
+ "tp": 0,
305
+ "fp": 0,
306
+ "fn": 723
307
+ },
308
+ "12": {
309
+ "f1_score": 0.0,
310
+ "precision": 0.0,
311
+ "recall": 0.0,
312
+ "tp": 0,
313
+ "fp": 0,
314
+ "fn": 725
315
+ },
316
+ "13": {
317
+ "f1_score": 0.0,
318
+ "precision": 0.0,
319
+ "recall": 0.0,
320
+ "tp": 0,
321
+ "fp": 0,
322
+ "fn": 721
323
+ },
324
+ "14": {
325
+ "f1_score": 0.0,
326
+ "precision": 0.0,
327
+ "recall": 0.0,
328
+ "tp": 0,
329
+ "fp": 0,
330
+ "fn": 694
331
+ },
332
+ "15": {
333
+ "f1_score": 0.0,
334
+ "precision": 0.0,
335
+ "recall": 0.0,
336
+ "tp": 0,
337
+ "fp": 0,
338
+ "fn": 714
339
+ },
340
+ "16": {
341
+ "f1_score": 0.0,
342
+ "precision": 0.0,
343
+ "recall": 0.0,
344
+ "tp": 0,
345
+ "fp": 0,
346
+ "fn": 725
347
+ },
348
+ "17": {
349
+ "f1_score": 0.0,
350
+ "precision": 0.0,
351
+ "recall": 0.0,
352
+ "tp": 0,
353
+ "fp": 0,
354
+ "fn": 723
355
+ },
356
+ "18": {
357
+ "f1_score": 0.0,
358
+ "precision": 0.0,
359
+ "recall": 0.0,
360
+ "tp": 0,
361
+ "fp": 0,
362
+ "fn": 723
363
+ },
364
+ "19": {
365
+ "f1_score": 0.0,
366
+ "precision": 0.0,
367
+ "recall": 0.0,
368
+ "tp": 0,
369
+ "fp": 0,
370
+ "fn": 724
371
+ },
372
+ "2": {
373
+ "f1_score": 0.0,
374
+ "precision": 0.0,
375
+ "recall": 0.0,
376
+ "tp": 0,
377
+ "fp": 0,
378
+ "fn": 725
379
+ },
380
+ "20": {
381
+ "f1_score": 0.0,
382
+ "precision": 0.0,
383
+ "recall": 0.0,
384
+ "tp": 0,
385
+ "fp": 0,
386
+ "fn": 721
387
+ },
388
+ "21": {
389
+ "f1_score": 0.0,
390
+ "precision": 0.0,
391
+ "recall": 0.0,
392
+ "tp": 0,
393
+ "fp": 0,
394
+ "fn": 725
395
+ },
396
+ "22": {
397
+ "f1_score": 0.0,
398
+ "precision": 0.0,
399
+ "recall": 0.0,
400
+ "tp": 0,
401
+ "fp": 0,
402
+ "fn": 716
403
+ },
404
+ "3": {
405
+ "f1_score": 0.0,
406
+ "precision": 0.0,
407
+ "recall": 0.0,
408
+ "tp": 0,
409
+ "fp": 0,
410
+ "fn": 724
411
+ },
412
+ "4": {
413
+ "f1_score": 0.0,
414
+ "precision": 0.0,
415
+ "recall": 0.0,
416
+ "tp": 0,
417
+ "fp": 0,
418
+ "fn": 723
419
+ },
420
+ "5": {
421
+ "f1_score": 0.0,
422
+ "precision": 0.0,
423
+ "recall": 0.0,
424
+ "tp": 0,
425
+ "fp": 0,
426
+ "fn": 723
427
+ },
428
+ "6": {
429
+ "f1_score": 0.0,
430
+ "precision": 0.0,
431
+ "recall": 0.0,
432
+ "tp": 0,
433
+ "fp": 0,
434
+ "fn": 725
435
+ },
436
+ "7": {
437
+ "f1_score": 0.0,
438
+ "precision": 0.0,
439
+ "recall": 0.0,
440
+ "tp": 0,
441
+ "fp": 0,
442
+ "fn": 722
443
+ },
444
+ "8": {
445
+ "f1_score": 0.0,
446
+ "precision": 0.0,
447
+ "recall": 0.0,
448
+ "tp": 0,
449
+ "fp": 0,
450
+ "fn": 720
451
+ },
452
+ "9": {
453
+ "f1_score": 0.0,
454
+ "precision": 0.0,
455
+ "recall": 0.0,
456
+ "tp": 0,
457
+ "fp": 0,
458
+ "fn": 724
459
+ },
460
+ "__unlabeled__": {
461
+ "f1_score": 0.22097087378640776,
462
+ "precision": 0.12505494505494505,
463
+ "recall": 0.9483333333333334,
464
+ "tp": 569,
465
+ "fp": 3981,
466
+ "fn": 31
467
+ },
468
+ "x": {
469
+ "f1_score": 0.0,
470
+ "precision": 0.0,
471
+ "recall": 0.0,
472
+ "tp": 0,
473
+ "fp": 0,
474
+ "fn": 558
475
+ },
476
+ "y": {
477
+ "f1_score": 0.0,
478
+ "precision": 0.0,
479
+ "recall": 0.0,
480
+ "tp": 0,
481
+ "fp": 0,
482
+ "fn": 174
483
+ }
484
+ },
485
+ "valid_samples": 961
486
+ },
487
+ "matching": {
488
+ "tp": 539,
489
+ "fp": 4011,
490
+ "fn": 16660
491
+ },
492
+ "iou_threshold": 0.5
493
+ },
494
+ "disease_diagnosis_classification": {
495
+ "metric": "balanced_accuracy",
496
+ "value": 0.7543050809989986,
497
+ "count": 2610,
498
+ "metrics": {
499
+ "balanced_accuracy": 0.7543050809989986,
500
+ "accuracy": 0.7513409961685824,
501
+ "f1_score": 0.7497215228330598,
502
+ "valid_samples": 2610
503
+ }
504
+ },
505
+ "multi_label_classification": {
506
+ "metric": "f1_score",
507
+ "value": 0.4437961099932931,
508
+ "count": 923,
509
+ "metrics": {
510
+ "f1_score": 0.4437961099932931,
511
+ "precision": 0.45774647887323944,
512
+ "recall": 0.45521849042975804,
513
+ "valid_samples": 923
514
+ }
515
+ },
516
+ "regression": {
517
+ "metric": "mean_absolute_error",
518
+ "value": 20.515590607695504,
519
+ "count": 202,
520
+ "metrics": {
521
+ "mean_absolute_error": 20.515590607695504,
522
+ "root_mean_squared_error": 35.72048318261565,
523
+ "valid_samples": 202
524
+ }
525
+ }
526
+ },
527
+ "num_rows": 4696,
528
+ "num_tasks": 4,
529
+ "predictions": "/scratch/atatc/output/medgemma-flare-2d-output/flare-medgemma-infer/testing_predictions.jsonl",
530
+ "split": "testing",
531
+ "model_variant": "predictions_file",
532
+ "num_predictions": 4696
533
+ },
534
+ "validation_public": {
535
+ "metrics": {
536
+ "cell_counting_mean_absolute_error": 275.64,
537
+ "cell_counting_root_mean_squared_error": 746.0054289346693,
538
+ "f1_iou_0.5": 0.20054200542005424,
539
+ "detection_f1_iou_0.5": 0.20054200542005424,
540
+ "detection_precision_iou_0.5": 0.20786516853932585,
541
+ "detection_recall_iou_0.5": 0.193717277486911,
542
+ "detection_f1_iou_0.3": 0.4281842818428184,
543
+ "detection_precision_iou_0.3": 0.4438202247191011,
544
+ "detection_recall_iou_0.3": 0.41361256544502617,
545
+ "detection_average_f1": 0.22547425474254745,
546
+ "detection_average_precision": 0.2337078651685393,
547
+ "detection_average_recall": 0.21780104712041887,
548
+ "detection_iou_0.3_f1_score": 0.4281842818428184,
549
+ "detection_iou_0.3_precision": 0.4438202247191011,
550
+ "detection_iou_0.3_recall": 0.41361256544502617,
551
+ "detection_iou_0.3_tp": 79,
552
+ "detection_iou_0.3_fp": 99,
553
+ "detection_iou_0.3_fn": 112,
554
+ "detection_iou_0.4_f1_score": 0.31978319783197834,
555
+ "detection_iou_0.4_precision": 0.33146067415730335,
556
+ "detection_iou_0.4_recall": 0.3089005235602094,
557
+ "detection_iou_0.4_tp": 59,
558
+ "detection_iou_0.4_fp": 119,
559
+ "detection_iou_0.4_fn": 132,
560
+ "detection_iou_0.5_f1_score": 0.20054200542005424,
561
+ "detection_iou_0.5_precision": 0.20786516853932585,
562
+ "detection_iou_0.5_recall": 0.193717277486911,
563
+ "detection_iou_0.5_tp": 37,
564
+ "detection_iou_0.5_fp": 141,
565
+ "detection_iou_0.5_fn": 154,
566
+ "detection_iou_0.6_f1_score": 0.11924119241192412,
567
+ "detection_iou_0.6_precision": 0.12359550561797752,
568
+ "detection_iou_0.6_recall": 0.11518324607329843,
569
+ "detection_iou_0.6_tp": 22,
570
+ "detection_iou_0.6_fp": 156,
571
+ "detection_iou_0.6_fn": 169,
572
+ "detection_iou_0.7_f1_score": 0.05962059620596206,
573
+ "detection_iou_0.7_precision": 0.06179775280898876,
574
+ "detection_iou_0.7_recall": 0.05759162303664921,
575
+ "detection_iou_0.7_tp": 11,
576
+ "detection_iou_0.7_fp": 167,
577
+ "detection_iou_0.7_fn": 180,
578
+ "balanced_accuracy": 0.5546660740446607,
579
+ "classification_accuracy": 0.7289484708839548,
580
+ "classification_f1_score": 0.7156102893269477,
581
+ "f1_score": 0.6075492977812565,
582
+ "multi_label_f1_score": 0.6075492977812565,
583
+ "multi_label_precision": 0.6396023564064801,
584
+ "multi_label_recall": 0.6242071674030437,
585
+ "green_score": 0.7304914922267108,
586
+ "crimson_score": 0.8560994341563786
587
+ },
588
+ "by_task": {
589
+ "cell_counting": {
590
+ "metric": "mean_absolute_error",
591
+ "value": 275.64,
592
+ "count": 100,
593
+ "metrics": {
594
+ "mean_absolute_error": 275.64,
595
+ "root_mean_squared_error": 746.0054289346693,
596
+ "valid_samples": 100
597
+ }
598
+ },
599
+ "detection": {
600
+ "metric": "f1_iou_0.5",
601
+ "value": 0.20054200542005424,
602
+ "count": 175,
603
+ "metrics": {
604
+ "f1_score": 0.20054200542005424,
605
+ "precision": 0.20786516853932585,
606
+ "recall": 0.193717277486911,
607
+ "f1_score_at_03": 0.4281842818428184,
608
+ "f1_score_at_05": 0.20054200542005424,
609
+ "average_f1": 0.22547425474254745,
610
+ "precision_at_03": 0.4438202247191011,
611
+ "recall_at_03": 0.41361256544502617,
612
+ "detailed_metrics": {
613
+ "IoU_0.3": {
614
+ "f1_score": 0.4281842818428184,
615
+ "precision": 0.4438202247191011,
616
+ "recall": 0.41361256544502617,
617
+ "tp": 79,
618
+ "fp": 99,
619
+ "fn": 112
620
+ },
621
+ "IoU_0.4": {
622
+ "f1_score": 0.31978319783197834,
623
+ "precision": 0.33146067415730335,
624
+ "recall": 0.3089005235602094,
625
+ "tp": 59,
626
+ "fp": 119,
627
+ "fn": 132
628
+ },
629
+ "IoU_0.5": {
630
+ "f1_score": 0.20054200542005424,
631
+ "precision": 0.20786516853932585,
632
+ "recall": 0.193717277486911,
633
+ "tp": 37,
634
+ "fp": 141,
635
+ "fn": 154
636
+ },
637
+ "IoU_0.6": {
638
+ "f1_score": 0.11924119241192412,
639
+ "precision": 0.12359550561797752,
640
+ "recall": 0.11518324607329843,
641
+ "tp": 22,
642
+ "fp": 156,
643
+ "fn": 169
644
+ },
645
+ "IoU_0.7": {
646
+ "f1_score": 0.05962059620596206,
647
+ "precision": 0.06179775280898876,
648
+ "recall": 0.05759162303664921,
649
+ "tp": 11,
650
+ "fp": 167,
651
+ "fn": 180
652
+ }
653
+ },
654
+ "coco_style_metrics": {
655
+ "average_f1": 0.22547425474254745,
656
+ "average_precision": 0.2337078651685393,
657
+ "average_recall": 0.21780104712041887
658
+ },
659
+ "per_chromosome_metrics": {
660
+ "__unlabeled__": {
661
+ "f1_score": 0.4281842818428184,
662
+ "precision": 0.4438202247191011,
663
+ "recall": 0.41361256544502617,
664
+ "tp": 79,
665
+ "fp": 99,
666
+ "fn": 112
667
+ }
668
+ },
669
+ "valid_samples": 175
670
+ },
671
+ "matching": {
672
+ "tp": 37,
673
+ "fp": 141,
674
+ "fn": 154
675
+ },
676
+ "iou_threshold": 0.5
677
+ },
678
+ "disease_diagnosis_classification": {
679
+ "metric": "balanced_accuracy",
680
+ "value": 0.5546660740446607,
681
+ "count": 2387,
682
+ "metrics": {
683
+ "balanced_accuracy": 0.5546660740446607,
684
+ "accuracy": 0.7289484708839548,
685
+ "f1_score": 0.7156102893269477,
686
+ "valid_samples": 2387
687
+ }
688
+ },
689
+ "multi_label_classification": {
690
+ "metric": "f1_score",
691
+ "value": 0.6075492977812565,
692
+ "count": 970,
693
+ "metrics": {
694
+ "f1_score": 0.6075492977812565,
695
+ "precision": 0.6396023564064801,
696
+ "recall": 0.6242071674030437,
697
+ "valid_samples": 970
698
+ }
699
+ },
700
+ "report_generation": {
701
+ "metric": "green_score",
702
+ "value": 0.7304914922267108,
703
+ "count": 1945,
704
+ "metrics": {
705
+ "green_score": 0.7304914922267108,
706
+ "valid_samples": 1945,
707
+ "crimson_score": 0.8560994341563786,
708
+ "crimson_valid_samples": 1944,
709
+ "crimson_failed_samples": 1,
710
+ "crimson_false_findings": 0.06532921810699588,
711
+ "crimson_missing_findings": 0.19753086419753085,
712
+ "crimson_attribute_errors": 0.07150205761316872,
713
+ "crimson_location_errors": 0.015432098765432098,
714
+ "crimson_severity_errors": 0.014917695473251029,
715
+ "crimson_descriptor_errors": 0.00977366255144033,
716
+ "crimson_measurement_errors": 0.0,
717
+ "crimson_certainty_errors": 0.00102880658436214,
718
+ "crimson_unspecific_errors": 0.018518518518518517,
719
+ "crimson_overinterpretation_errors": 0.0015432098765432098,
720
+ "crimson_temporal_errors": 0.010802469135802469,
721
+ "crimson_weighted_false_findings": 0.016075102880658436,
722
+ "crimson_weighted_missing_findings": 0.05066872427983539,
723
+ "crimson_weighted_attribute_errors": 0.00565843621399177,
724
+ "crimson_N_G": 0.06532921810699588,
725
+ "crimson_E_penalty": 0.016075102880658436,
726
+ "crimson_correct": 0.0122599451303155,
727
+ "crimson_errors_more_than_correct": 0.003815157750342936,
728
+ "crimson_S": 0.8190018943105363
729
+ }
730
+ }
731
+ },
732
+ "num_rows": 5577,
733
+ "num_tasks": 5,
734
+ "predictions": "/scratch/atatc/output/medgemma-flare-2d-output/flare-medgemma-infer/validation_public_predictions.jsonl",
735
+ "split": "validation_public",
736
+ "model_variant": "predictions_file",
737
+ "num_predictions": 5577
738
+ },
739
+ "validation_hidden": {
740
+ "metrics": {
741
+ "f1_iou_0.5": 0.02654867256637168,
742
+ "detection_f1_iou_0.5": 0.02654867256637168,
743
+ "detection_precision_iou_0.5": 0.06709741550695825,
744
+ "detection_recall_iou_0.5": 0.01654817357195391,
745
+ "detection_f1_iou_0.3": 0.10658800393313668,
746
+ "detection_precision_iou_0.3": 0.26938369781312127,
747
+ "detection_recall_iou_0.3": 0.06643785241480755,
748
+ "detection_average_f1": 0.04228121927236972,
749
+ "detection_average_precision": 0.10685884691848906,
750
+ "detection_average_recall": 0.026354498651630302,
751
+ "detection_iou_0.3_f1_score": 0.10658800393313668,
752
+ "detection_iou_0.3_precision": 0.26938369781312127,
753
+ "detection_iou_0.3_recall": 0.06643785241480755,
754
+ "detection_iou_0.3_tp": 542,
755
+ "detection_iou_0.3_fp": 1470,
756
+ "detection_iou_0.3_fn": 7616,
757
+ "detection_iou_0.4_f1_score": 0.05506391347099312,
758
+ "detection_iou_0.4_precision": 0.13916500994035785,
759
+ "detection_iou_0.4_recall": 0.03432213777886737,
760
+ "detection_iou_0.4_tp": 280,
761
+ "detection_iou_0.4_fp": 1732,
762
+ "detection_iou_0.4_fn": 7878,
763
+ "detection_iou_0.5_f1_score": 0.02654867256637168,
764
+ "detection_iou_0.5_precision": 0.06709741550695825,
765
+ "detection_iou_0.5_recall": 0.01654817357195391,
766
+ "detection_iou_0.5_tp": 135,
767
+ "detection_iou_0.5_fp": 1877,
768
+ "detection_iou_0.5_fn": 8023,
769
+ "detection_iou_0.6_f1_score": 0.014945919370698134,
770
+ "detection_iou_0.6_precision": 0.03777335984095427,
771
+ "detection_iou_0.6_recall": 0.009316008825692572,
772
+ "detection_iou_0.6_tp": 76,
773
+ "detection_iou_0.6_fp": 1936,
774
+ "detection_iou_0.6_fn": 8082,
775
+ "detection_iou_0.7_f1_score": 0.008259587020648967,
776
+ "detection_iou_0.7_precision": 0.020874751491053677,
777
+ "detection_iou_0.7_recall": 0.0051483206668301055,
778
+ "detection_iou_0.7_tp": 42,
779
+ "detection_iou_0.7_fp": 1970,
780
+ "detection_iou_0.7_fn": 8116,
781
+ "balanced_accuracy": 0.7831332298755335,
782
+ "classification_accuracy": 0.783303730017762,
783
+ "classification_f1_score": 0.780685616728517,
784
+ "f1_score": 0.44493361285814115,
785
+ "multi_label_f1_score": 0.44493361285814115,
786
+ "multi_label_precision": 0.4612159329140461,
787
+ "multi_label_recall": 0.45317959468902863,
788
+ "regression_mean_absolute_error": 15.5031279181712,
789
+ "regression_root_mean_squared_error": 20.465680477276777
790
+ },
791
+ "by_task": {
792
+ "detection": {
793
+ "metric": "f1_iou_0.5",
794
+ "value": 0.02654867256637168,
795
+ "count": 256,
796
+ "metrics": {
797
+ "f1_score": 0.02654867256637168,
798
+ "precision": 0.06709741550695825,
799
+ "recall": 0.01654817357195391,
800
+ "f1_score_at_03": 0.10658800393313668,
801
+ "f1_score_at_05": 0.02654867256637168,
802
+ "average_f1": 0.04228121927236972,
803
+ "precision_at_03": 0.26938369781312127,
804
+ "recall_at_03": 0.06643785241480755,
805
+ "detailed_metrics": {
806
+ "IoU_0.3": {
807
+ "f1_score": 0.10658800393313668,
808
+ "precision": 0.26938369781312127,
809
+ "recall": 0.06643785241480755,
810
+ "tp": 542,
811
+ "fp": 1470,
812
+ "fn": 7616
813
+ },
814
+ "IoU_0.4": {
815
+ "f1_score": 0.05506391347099312,
816
+ "precision": 0.13916500994035785,
817
+ "recall": 0.03432213777886737,
818
+ "tp": 280,
819
+ "fp": 1732,
820
+ "fn": 7878
821
+ },
822
+ "IoU_0.5": {
823
+ "f1_score": 0.02654867256637168,
824
+ "precision": 0.06709741550695825,
825
+ "recall": 0.01654817357195391,
826
+ "tp": 135,
827
+ "fp": 1877,
828
+ "fn": 8023
829
+ },
830
+ "IoU_0.6": {
831
+ "f1_score": 0.014945919370698134,
832
+ "precision": 0.03777335984095427,
833
+ "recall": 0.009316008825692572,
834
+ "tp": 76,
835
+ "fp": 1936,
836
+ "fn": 8082
837
+ },
838
+ "IoU_0.7": {
839
+ "f1_score": 0.008259587020648967,
840
+ "precision": 0.020874751491053677,
841
+ "recall": 0.0051483206668301055,
842
+ "tp": 42,
843
+ "fp": 1970,
844
+ "fn": 8116
845
+ }
846
+ },
847
+ "coco_style_metrics": {
848
+ "average_f1": 0.04228121927236972,
849
+ "average_precision": 0.10685884691848906,
850
+ "average_recall": 0.026354498651630302
851
+ },
852
+ "per_chromosome_metrics": {
853
+ "1": {
854
+ "f1_score": 0.0,
855
+ "precision": 0.0,
856
+ "recall": 0.0,
857
+ "tp": 0,
858
+ "fp": 0,
859
+ "fn": 352
860
+ },
861
+ "10": {
862
+ "f1_score": 0.0,
863
+ "precision": 0.0,
864
+ "recall": 0.0,
865
+ "tp": 0,
866
+ "fp": 0,
867
+ "fn": 352
868
+ },
869
+ "11": {
870
+ "f1_score": 0.0,
871
+ "precision": 0.0,
872
+ "recall": 0.0,
873
+ "tp": 0,
874
+ "fp": 0,
875
+ "fn": 353
876
+ },
877
+ "12": {
878
+ "f1_score": 0.0,
879
+ "precision": 0.0,
880
+ "recall": 0.0,
881
+ "tp": 0,
882
+ "fp": 0,
883
+ "fn": 351
884
+ },
885
+ "13": {
886
+ "f1_score": 0.0,
887
+ "precision": 0.0,
888
+ "recall": 0.0,
889
+ "tp": 0,
890
+ "fp": 0,
891
+ "fn": 349
892
+ },
893
+ "14": {
894
+ "f1_score": 0.0,
895
+ "precision": 0.0,
896
+ "recall": 0.0,
897
+ "tp": 0,
898
+ "fp": 0,
899
+ "fn": 340
900
+ },
901
+ "15": {
902
+ "f1_score": 0.0,
903
+ "precision": 0.0,
904
+ "recall": 0.0,
905
+ "tp": 0,
906
+ "fp": 0,
907
+ "fn": 351
908
+ },
909
+ "16": {
910
+ "f1_score": 0.0,
911
+ "precision": 0.0,
912
+ "recall": 0.0,
913
+ "tp": 0,
914
+ "fp": 0,
915
+ "fn": 352
916
+ },
917
+ "17": {
918
+ "f1_score": 0.0,
919
+ "precision": 0.0,
920
+ "recall": 0.0,
921
+ "tp": 0,
922
+ "fp": 0,
923
+ "fn": 350
924
+ },
925
+ "18": {
926
+ "f1_score": 0.0,
927
+ "precision": 0.0,
928
+ "recall": 0.0,
929
+ "tp": 0,
930
+ "fp": 0,
931
+ "fn": 353
932
+ },
933
+ "19": {
934
+ "f1_score": 0.0,
935
+ "precision": 0.0,
936
+ "recall": 0.0,
937
+ "tp": 0,
938
+ "fp": 0,
939
+ "fn": 352
940
+ },
941
+ "2": {
942
+ "f1_score": 0.0,
943
+ "precision": 0.0,
944
+ "recall": 0.0,
945
+ "tp": 0,
946
+ "fp": 0,
947
+ "fn": 352
948
+ },
949
+ "20": {
950
+ "f1_score": 0.0,
951
+ "precision": 0.0,
952
+ "recall": 0.0,
953
+ "tp": 0,
954
+ "fp": 0,
955
+ "fn": 354
956
+ },
957
+ "21": {
958
+ "f1_score": 0.0,
959
+ "precision": 0.0,
960
+ "recall": 0.0,
961
+ "tp": 0,
962
+ "fp": 0,
963
+ "fn": 357
964
+ },
965
+ "22": {
966
+ "f1_score": 0.0,
967
+ "precision": 0.0,
968
+ "recall": 0.0,
969
+ "tp": 0,
970
+ "fp": 0,
971
+ "fn": 350
972
+ },
973
+ "3": {
974
+ "f1_score": 0.0,
975
+ "precision": 0.0,
976
+ "recall": 0.0,
977
+ "tp": 0,
978
+ "fp": 0,
979
+ "fn": 350
980
+ },
981
+ "4": {
982
+ "f1_score": 0.0,
983
+ "precision": 0.0,
984
+ "recall": 0.0,
985
+ "tp": 0,
986
+ "fp": 0,
987
+ "fn": 351
988
+ },
989
+ "5": {
990
+ "f1_score": 0.0,
991
+ "precision": 0.0,
992
+ "recall": 0.0,
993
+ "tp": 0,
994
+ "fp": 0,
995
+ "fn": 350
996
+ },
997
+ "6": {
998
+ "f1_score": 0.0,
999
+ "precision": 0.0,
1000
+ "recall": 0.0,
1001
+ "tp": 0,
1002
+ "fp": 0,
1003
+ "fn": 353
1004
+ },
1005
+ "7": {
1006
+ "f1_score": 0.0,
1007
+ "precision": 0.0,
1008
+ "recall": 0.0,
1009
+ "tp": 0,
1010
+ "fp": 0,
1011
+ "fn": 353
1012
+ },
1013
+ "8": {
1014
+ "f1_score": 0.0,
1015
+ "precision": 0.0,
1016
+ "recall": 0.0,
1017
+ "tp": 0,
1018
+ "fp": 0,
1019
+ "fn": 351
1020
+ },
1021
+ "9": {
1022
+ "f1_score": 0.0,
1023
+ "precision": 0.0,
1024
+ "recall": 0.0,
1025
+ "tp": 0,
1026
+ "fp": 0,
1027
+ "fn": 349
1028
+ },
1029
+ "__unlabeled__": {
1030
+ "f1_score": 0.07361376673040154,
1031
+ "precision": 0.03827037773359841,
1032
+ "recall": 0.9625,
1033
+ "tp": 77,
1034
+ "fp": 1935,
1035
+ "fn": 3
1036
+ },
1037
+ "x": {
1038
+ "f1_score": 0.0,
1039
+ "precision": 0.0,
1040
+ "recall": 0.0,
1041
+ "tp": 0,
1042
+ "fp": 0,
1043
+ "fn": 277
1044
+ },
1045
+ "y": {
1046
+ "f1_score": 0.0,
1047
+ "precision": 0.0,
1048
+ "recall": 0.0,
1049
+ "tp": 0,
1050
+ "fp": 0,
1051
+ "fn": 76
1052
+ }
1053
+ },
1054
+ "valid_samples": 256
1055
+ },
1056
+ "matching": {
1057
+ "tp": 135,
1058
+ "fp": 1877,
1059
+ "fn": 8023
1060
+ },
1061
+ "iou_threshold": 0.5
1062
+ },
1063
+ "disease_diagnosis_classification": {
1064
+ "metric": "balanced_accuracy",
1065
+ "value": 0.7831332298755335,
1066
+ "count": 1126,
1067
+ "metrics": {
1068
+ "balanced_accuracy": 0.7831332298755335,
1069
+ "accuracy": 0.783303730017762,
1070
+ "f1_score": 0.780685616728517,
1071
+ "valid_samples": 1126
1072
+ }
1073
+ },
1074
+ "multi_label_classification": {
1075
+ "metric": "f1_score",
1076
+ "value": 0.44493361285814115,
1077
+ "count": 477,
1078
+ "metrics": {
1079
+ "f1_score": 0.44493361285814115,
1080
+ "precision": 0.4612159329140461,
1081
+ "recall": 0.45317959468902863,
1082
+ "valid_samples": 477
1083
+ }
1084
+ },
1085
+ "regression": {
1086
+ "metric": "mean_absolute_error",
1087
+ "value": 15.5031279181712,
1088
+ "count": 100,
1089
+ "metrics": {
1090
+ "mean_absolute_error": 15.5031279181712,
1091
+ "root_mean_squared_error": 20.465680477276777,
1092
+ "valid_samples": 100
1093
+ }
1094
+ }
1095
+ },
1096
+ "num_rows": 1959,
1097
+ "num_tasks": 4,
1098
+ "predictions": "/scratch/atatc/output/medgemma-flare-2d-output/flare-medgemma-infer/validation_hidden_predictions.jsonl",
1099
+ "split": "validation_hidden",
1100
+ "model_variant": "predictions_file",
1101
+ "num_predictions": 3918
1102
+ }
1103
+ },
1104
+ "num_rows": 12232,
1105
+ "num_splits": 3,
1106
+ "num_tasks": 6,
1107
+ "splits": [
1108
+ "testing",
1109
+ "validation_public",
1110
+ "validation_hidden"
1111
+ ],
1112
+ "model_variant": "predictions_file"
1113
+ }
flare-medgemma-eval/scores.json ADDED
@@ -0,0 +1,214 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "mean": {
3
+ "balanced_accuracy": 0.6973681283063976,
4
+ "cell_counting_mean_absolute_error": 275.64,
5
+ "cell_counting_root_mean_squared_error": 746.0054289346693,
6
+ "classification_accuracy": 0.7545310656900996,
7
+ "classification_f1_score": 0.7486724762961748,
8
+ "crimson_score": 0.8560994341563786,
9
+ "detection_average_f1": 0.11028574193986596,
10
+ "detection_average_precision": 0.16379329963340505,
11
+ "detection_average_recall": 0.09468440474704874,
12
+ "detection_f1_iou_0.3": 0.22629028834032594,
13
+ "detection_f1_iou_0.5": 0.09221872508355598,
14
+ "detection_iou_0.3_f1_score": 0.22629028834032594,
15
+ "detection_iou_0.3_fn": 7786.666666666667,
16
+ "detection_iou_0.3_fp": 1517.3333333333333,
17
+ "detection_iou_0.3_precision": 0.35253317564260894,
18
+ "detection_iou_0.3_recall": 0.19038678870421302,
19
+ "detection_iou_0.3_tp": 729.3333333333334,
20
+ "detection_iou_0.4_f1_score": 0.1537013168992954,
21
+ "detection_iou_0.4_fn": 8090.333333333333,
22
+ "detection_iou_0.4_fp": 1821.0,
23
+ "detection_iou_0.4_precision": 0.22559317675050247,
24
+ "detection_iou_0.4_recall": 0.13258690529237713,
25
+ "detection_iou_0.4_tp": 425.6666666666667,
26
+ "detection_iou_0.5_f1_score": 0.09221872508355598,
27
+ "detection_iou_0.5_fn": 8279.0,
28
+ "detection_iou_0.5_fp": 2009.6666666666667,
29
+ "detection_iou_0.5_precision": 0.13114137416927418,
30
+ "detection_iou_0.5_recall": 0.08053482746596542,
31
+ "detection_iou_0.5_tp": 237.0,
32
+ "detection_iou_0.6_f1_score": 0.05309723809769417,
33
+ "detection_iou_0.6_fn": 8392.333333333334,
34
+ "detection_iou_0.6_fp": 2123.0,
35
+ "detection_iou_0.6_precision": 0.07378962181964394,
36
+ "detection_iou_0.6_recall": 0.04679075692400229,
37
+ "detection_iou_0.6_tp": 123.66666666666667,
38
+ "detection_iou_0.7_f1_score": 0.026121141278458215,
39
+ "detection_iou_0.7_fn": 8460.333333333334,
40
+ "detection_iou_0.7_fp": 2191.0,
41
+ "detection_iou_0.7_precision": 0.03590914978499583,
42
+ "detection_iou_0.7_recall": 0.023122745348685792,
43
+ "detection_iou_0.7_tp": 55.666666666666664,
44
+ "detection_precision_iou_0.3": 0.35253317564260894,
45
+ "detection_precision_iou_0.5": 0.13114137416927418,
46
+ "detection_recall_iou_0.3": 0.19038678870421302,
47
+ "detection_recall_iou_0.5": 0.08053482746596542,
48
+ "f1_iou_0.5": 0.09221872508355598,
49
+ "f1_score": 0.4987596735442303,
50
+ "green_score": 0.7304914922267108,
51
+ "multi_label_f1_score": 0.4987596735442303,
52
+ "multi_label_precision": 0.5195215893979218,
53
+ "multi_label_recall": 0.5108684175072767,
54
+ "regression_mean_absolute_error": 18.00935926293335,
55
+ "regression_root_mean_squared_error": 28.093081829946215
56
+ },
57
+ "per_split": {
58
+ "testing": {
59
+ "f1_iou_0.5": 0.04956549726424204,
60
+ "detection_f1_iou_0.5": 0.04956549726424204,
61
+ "detection_precision_iou_0.5": 0.11846153846153847,
62
+ "detection_recall_iou_0.5": 0.03133903133903134,
63
+ "detection_f1_iou_0.3": 0.14409857924502276,
64
+ "detection_precision_iou_0.3": 0.3443956043956044,
65
+ "detection_recall_iou_0.3": 0.0911099482528054,
66
+ "detection_average_f1": 0.06310175180468068,
67
+ "detection_average_precision": 0.1508131868131868,
68
+ "detection_average_recall": 0.03989766846909704,
69
+ "detection_iou_0.3_f1_score": 0.14409857924502276,
70
+ "detection_iou_0.3_precision": 0.3443956043956044,
71
+ "detection_iou_0.3_recall": 0.0911099482528054,
72
+ "detection_iou_0.3_tp": 1567,
73
+ "detection_iou_0.3_fp": 2983,
74
+ "detection_iou_0.3_fn": 15632,
75
+ "detection_iou_0.4_f1_score": 0.08625683939491473,
76
+ "detection_iou_0.4_precision": 0.20615384615384616,
77
+ "detection_iou_0.4_recall": 0.05453805453805454,
78
+ "detection_iou_0.4_tp": 938,
79
+ "detection_iou_0.4_fp": 3612,
80
+ "detection_iou_0.4_fn": 16261,
81
+ "detection_iou_0.5_f1_score": 0.04956549726424204,
82
+ "detection_iou_0.5_precision": 0.11846153846153847,
83
+ "detection_iou_0.5_recall": 0.03133903133903134,
84
+ "detection_iou_0.5_tp": 539,
85
+ "detection_iou_0.5_fp": 4011,
86
+ "detection_iou_0.5_fn": 16660,
87
+ "detection_iou_0.6_f1_score": 0.02510460251046025,
88
+ "detection_iou_0.6_precision": 0.06,
89
+ "detection_iou_0.6_recall": 0.015873015873015872,
90
+ "detection_iou_0.6_tp": 273,
91
+ "detection_iou_0.6_fp": 4277,
92
+ "detection_iou_0.6_fn": 16926,
93
+ "detection_iou_0.7_f1_score": 0.01048324060876362,
94
+ "detection_iou_0.7_precision": 0.025054945054945054,
95
+ "detection_iou_0.7_recall": 0.006628292342578057,
96
+ "detection_iou_0.7_tp": 114,
97
+ "detection_iou_0.7_fp": 4436,
98
+ "detection_iou_0.7_fn": 17085,
99
+ "balanced_accuracy": 0.7543050809989986,
100
+ "classification_accuracy": 0.7513409961685824,
101
+ "classification_f1_score": 0.7497215228330598,
102
+ "f1_score": 0.4437961099932931,
103
+ "multi_label_f1_score": 0.4437961099932931,
104
+ "multi_label_precision": 0.45774647887323944,
105
+ "multi_label_recall": 0.45521849042975804,
106
+ "regression_mean_absolute_error": 20.515590607695504,
107
+ "regression_root_mean_squared_error": 35.72048318261565
108
+ },
109
+ "validation_public": {
110
+ "cell_counting_mean_absolute_error": 275.64,
111
+ "cell_counting_root_mean_squared_error": 746.0054289346693,
112
+ "f1_iou_0.5": 0.20054200542005424,
113
+ "detection_f1_iou_0.5": 0.20054200542005424,
114
+ "detection_precision_iou_0.5": 0.20786516853932585,
115
+ "detection_recall_iou_0.5": 0.193717277486911,
116
+ "detection_f1_iou_0.3": 0.4281842818428184,
117
+ "detection_precision_iou_0.3": 0.4438202247191011,
118
+ "detection_recall_iou_0.3": 0.41361256544502617,
119
+ "detection_average_f1": 0.22547425474254745,
120
+ "detection_average_precision": 0.2337078651685393,
121
+ "detection_average_recall": 0.21780104712041887,
122
+ "detection_iou_0.3_f1_score": 0.4281842818428184,
123
+ "detection_iou_0.3_precision": 0.4438202247191011,
124
+ "detection_iou_0.3_recall": 0.41361256544502617,
125
+ "detection_iou_0.3_tp": 79,
126
+ "detection_iou_0.3_fp": 99,
127
+ "detection_iou_0.3_fn": 112,
128
+ "detection_iou_0.4_f1_score": 0.31978319783197834,
129
+ "detection_iou_0.4_precision": 0.33146067415730335,
130
+ "detection_iou_0.4_recall": 0.3089005235602094,
131
+ "detection_iou_0.4_tp": 59,
132
+ "detection_iou_0.4_fp": 119,
133
+ "detection_iou_0.4_fn": 132,
134
+ "detection_iou_0.5_f1_score": 0.20054200542005424,
135
+ "detection_iou_0.5_precision": 0.20786516853932585,
136
+ "detection_iou_0.5_recall": 0.193717277486911,
137
+ "detection_iou_0.5_tp": 37,
138
+ "detection_iou_0.5_fp": 141,
139
+ "detection_iou_0.5_fn": 154,
140
+ "detection_iou_0.6_f1_score": 0.11924119241192412,
141
+ "detection_iou_0.6_precision": 0.12359550561797752,
142
+ "detection_iou_0.6_recall": 0.11518324607329843,
143
+ "detection_iou_0.6_tp": 22,
144
+ "detection_iou_0.6_fp": 156,
145
+ "detection_iou_0.6_fn": 169,
146
+ "detection_iou_0.7_f1_score": 0.05962059620596206,
147
+ "detection_iou_0.7_precision": 0.06179775280898876,
148
+ "detection_iou_0.7_recall": 0.05759162303664921,
149
+ "detection_iou_0.7_tp": 11,
150
+ "detection_iou_0.7_fp": 167,
151
+ "detection_iou_0.7_fn": 180,
152
+ "balanced_accuracy": 0.5546660740446607,
153
+ "classification_accuracy": 0.7289484708839548,
154
+ "classification_f1_score": 0.7156102893269477,
155
+ "f1_score": 0.6075492977812565,
156
+ "multi_label_f1_score": 0.6075492977812565,
157
+ "multi_label_precision": 0.6396023564064801,
158
+ "multi_label_recall": 0.6242071674030437,
159
+ "green_score": 0.7304914922267108,
160
+ "crimson_score": 0.8560994341563786
161
+ },
162
+ "validation_hidden": {
163
+ "f1_iou_0.5": 0.02654867256637168,
164
+ "detection_f1_iou_0.5": 0.02654867256637168,
165
+ "detection_precision_iou_0.5": 0.06709741550695825,
166
+ "detection_recall_iou_0.5": 0.01654817357195391,
167
+ "detection_f1_iou_0.3": 0.10658800393313668,
168
+ "detection_precision_iou_0.3": 0.26938369781312127,
169
+ "detection_recall_iou_0.3": 0.06643785241480755,
170
+ "detection_average_f1": 0.04228121927236972,
171
+ "detection_average_precision": 0.10685884691848906,
172
+ "detection_average_recall": 0.026354498651630302,
173
+ "detection_iou_0.3_f1_score": 0.10658800393313668,
174
+ "detection_iou_0.3_precision": 0.26938369781312127,
175
+ "detection_iou_0.3_recall": 0.06643785241480755,
176
+ "detection_iou_0.3_tp": 542,
177
+ "detection_iou_0.3_fp": 1470,
178
+ "detection_iou_0.3_fn": 7616,
179
+ "detection_iou_0.4_f1_score": 0.05506391347099312,
180
+ "detection_iou_0.4_precision": 0.13916500994035785,
181
+ "detection_iou_0.4_recall": 0.03432213777886737,
182
+ "detection_iou_0.4_tp": 280,
183
+ "detection_iou_0.4_fp": 1732,
184
+ "detection_iou_0.4_fn": 7878,
185
+ "detection_iou_0.5_f1_score": 0.02654867256637168,
186
+ "detection_iou_0.5_precision": 0.06709741550695825,
187
+ "detection_iou_0.5_recall": 0.01654817357195391,
188
+ "detection_iou_0.5_tp": 135,
189
+ "detection_iou_0.5_fp": 1877,
190
+ "detection_iou_0.5_fn": 8023,
191
+ "detection_iou_0.6_f1_score": 0.014945919370698134,
192
+ "detection_iou_0.6_precision": 0.03777335984095427,
193
+ "detection_iou_0.6_recall": 0.009316008825692572,
194
+ "detection_iou_0.6_tp": 76,
195
+ "detection_iou_0.6_fp": 1936,
196
+ "detection_iou_0.6_fn": 8082,
197
+ "detection_iou_0.7_f1_score": 0.008259587020648967,
198
+ "detection_iou_0.7_precision": 0.020874751491053677,
199
+ "detection_iou_0.7_recall": 0.0051483206668301055,
200
+ "detection_iou_0.7_tp": 42,
201
+ "detection_iou_0.7_fp": 1970,
202
+ "detection_iou_0.7_fn": 8116,
203
+ "balanced_accuracy": 0.7831332298755335,
204
+ "classification_accuracy": 0.783303730017762,
205
+ "classification_f1_score": 0.780685616728517,
206
+ "f1_score": 0.44493361285814115,
207
+ "multi_label_f1_score": 0.44493361285814115,
208
+ "multi_label_precision": 0.4612159329140461,
209
+ "multi_label_recall": 0.45317959468902863,
210
+ "regression_mean_absolute_error": 15.5031279181712,
211
+ "regression_root_mean_squared_error": 20.465680477276777
212
+ }
213
+ }
214
+ }
flare-medgemma-infer/inference_details.json ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "splits": [
3
+ "testing",
4
+ "validation_public",
5
+ "validation_hidden"
6
+ ],
7
+ "tasks": [
8
+ "disease_diagnosis_classification",
9
+ "multi_label_classification",
10
+ "detection",
11
+ "cell_counting",
12
+ "regression",
13
+ "report_generation"
14
+ ],
15
+ "model_variant": "adapter",
16
+ "split_results": {
17
+ "testing": {
18
+ "split": "testing",
19
+ "predictions": "/scratch/atatc/output/medgemma-flare-2d-output/flare-medgemma-infer/testing_predictions.jsonl",
20
+ "num_predictions": 4696,
21
+ "num_rows": 4696,
22
+ "model_variant": "adapter"
23
+ },
24
+ "validation_public": {
25
+ "split": "validation_public",
26
+ "predictions": "/scratch/atatc/output/medgemma-flare-2d-output/flare-medgemma-infer/validation_public_predictions.jsonl",
27
+ "num_predictions": 5577,
28
+ "num_rows": 5577,
29
+ "model_variant": "adapter"
30
+ },
31
+ "validation_hidden": {
32
+ "split": "validation_hidden",
33
+ "predictions": "/scratch/atatc/output/medgemma-flare-2d-output/flare-medgemma-infer/validation_hidden_predictions.jsonl",
34
+ "num_predictions": 3918,
35
+ "num_rows": 3918,
36
+ "model_variant": "adapter"
37
+ }
38
+ },
39
+ "num_predictions": 14191
40
+ }
flare-medgemma-infer/testing_predictions.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
flare-medgemma-infer/validation_hidden_predictions.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
flare-medgemma-infer/validation_public_predictions.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
flare-medgemma-medgemma15-lora/README.md ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: google/medgemma-1.5-4b-it
3
+ library_name: transformers
4
+ model_name: flare-medgemma-medgemma15-lora
5
+ tags:
6
+ - generated_from_trainer
7
+ - sft
8
+ - trl
9
+ licence: license
10
+ ---
11
+
12
+ # Model Card for flare-medgemma-medgemma15-lora
13
+
14
+ This model is a fine-tuned version of [google/medgemma-1.5-4b-it](https://huggingface.co/google/medgemma-1.5-4b-it).
15
+ It has been trained using [TRL](https://github.com/huggingface/trl).
16
+
17
+ ## Quick start
18
+
19
+ ```python
20
+ from transformers import pipeline
21
+
22
+ question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
23
+ generator = pipeline("text-generation", model="None", device="cuda")
24
+ output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
25
+ print(output["generated_text"])
26
+ ```
27
+
28
+ ## Training procedure
29
+
30
+ [<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/projectneura/medgemma15-flare-mllm-2d/runs/by79naj9)
31
+
32
+
33
+
34
+ This model was trained with SFT.
35
+
36
+ ### Framework versions
37
+
38
+ - TRL: 1.2.0+computecanada
39
+ - Transformers: 5.6.2+computecanada
40
+ - Pytorch: 2.9.0+computecanada
41
+ - Datasets: 4.8.4+computecanada
42
+ - Tokenizers: 0.22.2+computecanada
43
+
44
+ ## Citations
45
+
46
+
47
+
48
+ Cite TRL as:
49
+
50
+ ```bibtex
51
+ @software{vonwerra2020trl,
52
+ title = {{TRL: Transformers Reinforcement Learning}},
53
+ author = {von Werra, Leandro and Belkada, Younes and Tunstall, Lewis and Beeching, Edward and Thrush, Tristan and Lambert, Nathan and Huang, Shengyi and Rasul, Kashif and Gallouédec, Quentin},
54
+ license = {Apache-2.0},
55
+ url = {https://github.com/huggingface/trl},
56
+ year = {2020}
57
+ }
58
+ ```
flare-medgemma-medgemma15-lora/checkpoint-17000/README.md ADDED
@@ -0,0 +1,209 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: google/medgemma-1.5-4b-it
3
+ library_name: peft
4
+ pipeline_tag: text-generation
5
+ tags:
6
+ - base_model:adapter:google/medgemma-1.5-4b-it
7
+ - lora
8
+ - sft
9
+ - transformers
10
+ - trl
11
+ ---
12
+
13
+ # Model Card for Model ID
14
+
15
+ <!-- Provide a quick summary of what the model is/does. -->
16
+
17
+
18
+
19
+ ## Model Details
20
+
21
+ ### Model Description
22
+
23
+ <!-- Provide a longer summary of what this model is. -->
24
+
25
+
26
+
27
+ - **Developed by:** [More Information Needed]
28
+ - **Funded by [optional]:** [More Information Needed]
29
+ - **Shared by [optional]:** [More Information Needed]
30
+ - **Model type:** [More Information Needed]
31
+ - **Language(s) (NLP):** [More Information Needed]
32
+ - **License:** [More Information Needed]
33
+ - **Finetuned from model [optional]:** [More Information Needed]
34
+
35
+ ### Model Sources [optional]
36
+
37
+ <!-- Provide the basic links for the model. -->
38
+
39
+ - **Repository:** [More Information Needed]
40
+ - **Paper [optional]:** [More Information Needed]
41
+ - **Demo [optional]:** [More Information Needed]
42
+
43
+ ## Uses
44
+
45
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
46
+
47
+ ### Direct Use
48
+
49
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
50
+
51
+ [More Information Needed]
52
+
53
+ ### Downstream Use [optional]
54
+
55
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
56
+
57
+ [More Information Needed]
58
+
59
+ ### Out-of-Scope Use
60
+
61
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
62
+
63
+ [More Information Needed]
64
+
65
+ ## Bias, Risks, and Limitations
66
+
67
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
68
+
69
+ [More Information Needed]
70
+
71
+ ### Recommendations
72
+
73
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
74
+
75
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
76
+
77
+ ## How to Get Started with the Model
78
+
79
+ Use the code below to get started with the model.
80
+
81
+ [More Information Needed]
82
+
83
+ ## Training Details
84
+
85
+ ### Training Data
86
+
87
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
88
+
89
+ [More Information Needed]
90
+
91
+ ### Training Procedure
92
+
93
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
94
+
95
+ #### Preprocessing [optional]
96
+
97
+ [More Information Needed]
98
+
99
+
100
+ #### Training Hyperparameters
101
+
102
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
103
+
104
+ #### Speeds, Sizes, Times [optional]
105
+
106
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
107
+
108
+ [More Information Needed]
109
+
110
+ ## Evaluation
111
+
112
+ <!-- This section describes the evaluation protocols and provides the results. -->
113
+
114
+ ### Testing Data, Factors & Metrics
115
+
116
+ #### Testing Data
117
+
118
+ <!-- This should link to a Dataset Card if possible. -->
119
+
120
+ [More Information Needed]
121
+
122
+ #### Factors
123
+
124
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
125
+
126
+ [More Information Needed]
127
+
128
+ #### Metrics
129
+
130
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
131
+
132
+ [More Information Needed]
133
+
134
+ ### Results
135
+
136
+ [More Information Needed]
137
+
138
+ #### Summary
139
+
140
+
141
+
142
+ ## Model Examination [optional]
143
+
144
+ <!-- Relevant interpretability work for the model goes here -->
145
+
146
+ [More Information Needed]
147
+
148
+ ## Environmental Impact
149
+
150
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
151
+
152
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
153
+
154
+ - **Hardware Type:** [More Information Needed]
155
+ - **Hours used:** [More Information Needed]
156
+ - **Cloud Provider:** [More Information Needed]
157
+ - **Compute Region:** [More Information Needed]
158
+ - **Carbon Emitted:** [More Information Needed]
159
+
160
+ ## Technical Specifications [optional]
161
+
162
+ ### Model Architecture and Objective
163
+
164
+ [More Information Needed]
165
+
166
+ ### Compute Infrastructure
167
+
168
+ [More Information Needed]
169
+
170
+ #### Hardware
171
+
172
+ [More Information Needed]
173
+
174
+ #### Software
175
+
176
+ [More Information Needed]
177
+
178
+ ## Citation [optional]
179
+
180
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
181
+
182
+ **BibTeX:**
183
+
184
+ [More Information Needed]
185
+
186
+ **APA:**
187
+
188
+ [More Information Needed]
189
+
190
+ ## Glossary [optional]
191
+
192
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
193
+
194
+ [More Information Needed]
195
+
196
+ ## More Information [optional]
197
+
198
+ [More Information Needed]
199
+
200
+ ## Model Card Authors [optional]
201
+
202
+ [More Information Needed]
203
+
204
+ ## Model Card Contact
205
+
206
+ [More Information Needed]
207
+ ### Framework versions
208
+
209
+ - PEFT 0.19.1
flare-medgemma-medgemma15-lora/checkpoint-17000/adapter_config.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alora_invocation_tokens": null,
3
+ "alpha_pattern": {},
4
+ "arrow_config": null,
5
+ "auto_mapping": null,
6
+ "base_model_name_or_path": "google/medgemma-1.5-4b-it",
7
+ "bias": "none",
8
+ "corda_config": null,
9
+ "ensure_weight_tying": false,
10
+ "eva_config": null,
11
+ "exclude_modules": null,
12
+ "fan_in_fan_out": false,
13
+ "inference_mode": true,
14
+ "init_lora_weights": true,
15
+ "layer_replication": null,
16
+ "layers_pattern": null,
17
+ "layers_to_transform": null,
18
+ "loftq_config": {},
19
+ "lora_alpha": 16,
20
+ "lora_bias": false,
21
+ "lora_dropout": 0.05,
22
+ "lora_ga_config": null,
23
+ "megatron_config": null,
24
+ "megatron_core": "megatron.core",
25
+ "modules_to_save": null,
26
+ "peft_type": "LORA",
27
+ "peft_version": "0.19.1",
28
+ "qalora_group_size": 16,
29
+ "r": 16,
30
+ "rank_pattern": {},
31
+ "revision": null,
32
+ "target_modules": [
33
+ "gate_proj",
34
+ "down_proj",
35
+ "fc2",
36
+ "k_proj",
37
+ "fc1",
38
+ "out_proj",
39
+ "v_proj",
40
+ "up_proj",
41
+ "o_proj",
42
+ "q_proj"
43
+ ],
44
+ "target_parameters": null,
45
+ "task_type": "CAUSAL_LM",
46
+ "trainable_token_indices": null,
47
+ "use_bdlora": null,
48
+ "use_dora": false,
49
+ "use_qalora": false,
50
+ "use_rslora": false
51
+ }
flare-medgemma-medgemma15-lora/checkpoint-17000/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5deacc63e2e19e06dc6fa23ba9f314b7e9fdfd1c5b78662dfbe9b2df57156003
3
+ size 77116432
flare-medgemma-medgemma15-lora/checkpoint-17000/chat_template.jinja ADDED
@@ -0,0 +1,47 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {{ bos_token }}
2
+ {%- if messages[0]['role'] == 'system' -%}
3
+ {%- if messages[0]['content'] is string -%}
4
+ {%- set first_user_prefix = messages[0]['content'] + '
5
+
6
+ ' -%}
7
+ {%- else -%}
8
+ {%- set first_user_prefix = messages[0]['content'][0]['text'] + '
9
+
10
+ ' -%}
11
+ {%- endif -%}
12
+ {%- set loop_messages = messages[1:] -%}
13
+ {%- else -%}
14
+ {%- set first_user_prefix = "" -%}
15
+ {%- set loop_messages = messages -%}
16
+ {%- endif -%}
17
+ {%- for message in loop_messages -%}
18
+ {%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) -%}
19
+ {{ raise_exception("Conversation roles must alternate user/assistant/user/assistant/...") }}
20
+ {%- endif -%}
21
+ {%- if (message['role'] == 'assistant') -%}
22
+ {%- set role = "model" -%}
23
+ {%- else -%}
24
+ {%- set role = message['role'] -%}
25
+ {%- endif -%}
26
+ {{ '<start_of_turn>' + role + '
27
+ ' + (first_user_prefix if loop.first else "") }}
28
+ {%- if message['content'] is string -%}
29
+ {{ message['content'] | trim }}
30
+ {%- elif message['content'] is iterable -%}
31
+ {%- for item in message['content'] -%}
32
+ {%- if item['type'] == 'image' -%}
33
+ {{ '<start_of_image>' }}
34
+ {%- elif item['type'] == 'text' -%}
35
+ {{ item['text'] | trim }}
36
+ {%- endif -%}
37
+ {%- endfor -%}
38
+ {%- else -%}
39
+ {{ raise_exception("Invalid content type") }}
40
+ {%- endif -%}
41
+ {{ '<end_of_turn>
42
+ ' }}
43
+ {%- endfor -%}
44
+ {%- if add_generation_prompt -%}
45
+ {{'<start_of_turn>model
46
+ '}}
47
+ {%- endif -%}
flare-medgemma-medgemma15-lora/checkpoint-17000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7559ba771bc087526dbdef612bc3b4aaa25f29bc246c08cc6b9672c6334e0a6d
3
+ size 79126053
flare-medgemma-medgemma15-lora/checkpoint-17000/processor_config.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "image_processor": {
3
+ "do_convert_rgb": true,
4
+ "do_normalize": true,
5
+ "do_rescale": true,
6
+ "do_resize": true,
7
+ "image_mean": [
8
+ 0.5,
9
+ 0.5,
10
+ 0.5
11
+ ],
12
+ "image_processor_type": "Gemma3ImageProcessor",
13
+ "image_seq_length": 256,
14
+ "image_std": [
15
+ 0.5,
16
+ 0.5,
17
+ 0.5
18
+ ],
19
+ "resample": 2,
20
+ "rescale_factor": 0.00392156862745098,
21
+ "size": {
22
+ "height": 896,
23
+ "width": 896
24
+ }
25
+ },
26
+ "image_seq_length": 256,
27
+ "processor_class": "Gemma3Processor"
28
+ }
flare-medgemma-medgemma15-lora/checkpoint-17000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0a1c79ea61d0fe0288a760b836bc3a6faeff6409e970cbc744fc9ea5d070775f
3
+ size 14645
flare-medgemma-medgemma15-lora/checkpoint-17000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f2dfb83f80df6a389cda9ac22132713fa71d2f7502990b7b8bb2b5cc52747294
3
+ size 1465
flare-medgemma-medgemma15-lora/checkpoint-17000/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:daab2354f8a74e70d70b4d1f804939b68a8c9624dd06cb7858e52dd8970e9726
3
+ size 33384567
flare-medgemma-medgemma15-lora/checkpoint-17000/tokenizer_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "backend": "tokenizers",
3
+ "boi_token": "<start_of_image>",
4
+ "bos_token": "<bos>",
5
+ "clean_up_tokenization_spaces": false,
6
+ "eoi_token": "<end_of_image>",
7
+ "eos_token": "<eos>",
8
+ "image_token": "<image_soft_token>",
9
+ "is_local": false,
10
+ "local_files_only": false,
11
+ "mask_token": "<mask>",
12
+ "model_max_length": 1000000000000000019884624838656,
13
+ "model_specific_special_tokens": {
14
+ "boi_token": "<start_of_image>",
15
+ "eoi_token": "<end_of_image>",
16
+ "image_token": "<image_soft_token>"
17
+ },
18
+ "pad_token": "<pad>",
19
+ "padding_side": "right",
20
+ "processor_class": "Gemma3Processor",
21
+ "sp_model_kwargs": null,
22
+ "spaces_between_special_tokens": false,
23
+ "tokenizer_class": "GemmaTokenizer",
24
+ "unk_token": "<unk>",
25
+ "use_default_system_prompt": false
26
+ }
flare-medgemma-medgemma15-lora/checkpoint-17000/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
flare-medgemma-medgemma15-lora/checkpoint-17000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e7cafb5cde52b908ce007ba70fb2b52cc4f4dffde9f907f7e5805f932473d809
3
+ size 5905
flare-medgemma-medgemma15-lora/checkpoint-17200/README.md ADDED
@@ -0,0 +1,209 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: google/medgemma-1.5-4b-it
3
+ library_name: peft
4
+ pipeline_tag: text-generation
5
+ tags:
6
+ - base_model:adapter:google/medgemma-1.5-4b-it
7
+ - lora
8
+ - sft
9
+ - transformers
10
+ - trl
11
+ ---
12
+
13
+ # Model Card for Model ID
14
+
15
+ <!-- Provide a quick summary of what the model is/does. -->
16
+
17
+
18
+
19
+ ## Model Details
20
+
21
+ ### Model Description
22
+
23
+ <!-- Provide a longer summary of what this model is. -->
24
+
25
+
26
+
27
+ - **Developed by:** [More Information Needed]
28
+ - **Funded by [optional]:** [More Information Needed]
29
+ - **Shared by [optional]:** [More Information Needed]
30
+ - **Model type:** [More Information Needed]
31
+ - **Language(s) (NLP):** [More Information Needed]
32
+ - **License:** [More Information Needed]
33
+ - **Finetuned from model [optional]:** [More Information Needed]
34
+
35
+ ### Model Sources [optional]
36
+
37
+ <!-- Provide the basic links for the model. -->
38
+
39
+ - **Repository:** [More Information Needed]
40
+ - **Paper [optional]:** [More Information Needed]
41
+ - **Demo [optional]:** [More Information Needed]
42
+
43
+ ## Uses
44
+
45
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
46
+
47
+ ### Direct Use
48
+
49
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
50
+
51
+ [More Information Needed]
52
+
53
+ ### Downstream Use [optional]
54
+
55
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
56
+
57
+ [More Information Needed]
58
+
59
+ ### Out-of-Scope Use
60
+
61
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
62
+
63
+ [More Information Needed]
64
+
65
+ ## Bias, Risks, and Limitations
66
+
67
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
68
+
69
+ [More Information Needed]
70
+
71
+ ### Recommendations
72
+
73
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
74
+
75
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
76
+
77
+ ## How to Get Started with the Model
78
+
79
+ Use the code below to get started with the model.
80
+
81
+ [More Information Needed]
82
+
83
+ ## Training Details
84
+
85
+ ### Training Data
86
+
87
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
88
+
89
+ [More Information Needed]
90
+
91
+ ### Training Procedure
92
+
93
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
94
+
95
+ #### Preprocessing [optional]
96
+
97
+ [More Information Needed]
98
+
99
+
100
+ #### Training Hyperparameters
101
+
102
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
103
+
104
+ #### Speeds, Sizes, Times [optional]
105
+
106
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
107
+
108
+ [More Information Needed]
109
+
110
+ ## Evaluation
111
+
112
+ <!-- This section describes the evaluation protocols and provides the results. -->
113
+
114
+ ### Testing Data, Factors & Metrics
115
+
116
+ #### Testing Data
117
+
118
+ <!-- This should link to a Dataset Card if possible. -->
119
+
120
+ [More Information Needed]
121
+
122
+ #### Factors
123
+
124
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
125
+
126
+ [More Information Needed]
127
+
128
+ #### Metrics
129
+
130
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
131
+
132
+ [More Information Needed]
133
+
134
+ ### Results
135
+
136
+ [More Information Needed]
137
+
138
+ #### Summary
139
+
140
+
141
+
142
+ ## Model Examination [optional]
143
+
144
+ <!-- Relevant interpretability work for the model goes here -->
145
+
146
+ [More Information Needed]
147
+
148
+ ## Environmental Impact
149
+
150
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
151
+
152
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
153
+
154
+ - **Hardware Type:** [More Information Needed]
155
+ - **Hours used:** [More Information Needed]
156
+ - **Cloud Provider:** [More Information Needed]
157
+ - **Compute Region:** [More Information Needed]
158
+ - **Carbon Emitted:** [More Information Needed]
159
+
160
+ ## Technical Specifications [optional]
161
+
162
+ ### Model Architecture and Objective
163
+
164
+ [More Information Needed]
165
+
166
+ ### Compute Infrastructure
167
+
168
+ [More Information Needed]
169
+
170
+ #### Hardware
171
+
172
+ [More Information Needed]
173
+
174
+ #### Software
175
+
176
+ [More Information Needed]
177
+
178
+ ## Citation [optional]
179
+
180
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
181
+
182
+ **BibTeX:**
183
+
184
+ [More Information Needed]
185
+
186
+ **APA:**
187
+
188
+ [More Information Needed]
189
+
190
+ ## Glossary [optional]
191
+
192
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
193
+
194
+ [More Information Needed]
195
+
196
+ ## More Information [optional]
197
+
198
+ [More Information Needed]
199
+
200
+ ## Model Card Authors [optional]
201
+
202
+ [More Information Needed]
203
+
204
+ ## Model Card Contact
205
+
206
+ [More Information Needed]
207
+ ### Framework versions
208
+
209
+ - PEFT 0.19.1
flare-medgemma-medgemma15-lora/checkpoint-17200/adapter_config.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alora_invocation_tokens": null,
3
+ "alpha_pattern": {},
4
+ "arrow_config": null,
5
+ "auto_mapping": null,
6
+ "base_model_name_or_path": "google/medgemma-1.5-4b-it",
7
+ "bias": "none",
8
+ "corda_config": null,
9
+ "ensure_weight_tying": false,
10
+ "eva_config": null,
11
+ "exclude_modules": null,
12
+ "fan_in_fan_out": false,
13
+ "inference_mode": true,
14
+ "init_lora_weights": true,
15
+ "layer_replication": null,
16
+ "layers_pattern": null,
17
+ "layers_to_transform": null,
18
+ "loftq_config": {},
19
+ "lora_alpha": 16,
20
+ "lora_bias": false,
21
+ "lora_dropout": 0.05,
22
+ "lora_ga_config": null,
23
+ "megatron_config": null,
24
+ "megatron_core": "megatron.core",
25
+ "modules_to_save": null,
26
+ "peft_type": "LORA",
27
+ "peft_version": "0.19.1",
28
+ "qalora_group_size": 16,
29
+ "r": 16,
30
+ "rank_pattern": {},
31
+ "revision": null,
32
+ "target_modules": [
33
+ "gate_proj",
34
+ "down_proj",
35
+ "fc2",
36
+ "k_proj",
37
+ "fc1",
38
+ "out_proj",
39
+ "v_proj",
40
+ "up_proj",
41
+ "o_proj",
42
+ "q_proj"
43
+ ],
44
+ "target_parameters": null,
45
+ "task_type": "CAUSAL_LM",
46
+ "trainable_token_indices": null,
47
+ "use_bdlora": null,
48
+ "use_dora": false,
49
+ "use_qalora": false,
50
+ "use_rslora": false
51
+ }
flare-medgemma-medgemma15-lora/checkpoint-17200/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d9a22fbb36edcf31dc600631dd9a152589f79fb5a576f680aa042ea7c5cd3008
3
+ size 77116432
flare-medgemma-medgemma15-lora/checkpoint-17200/chat_template.jinja ADDED
@@ -0,0 +1,47 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {{ bos_token }}
2
+ {%- if messages[0]['role'] == 'system' -%}
3
+ {%- if messages[0]['content'] is string -%}
4
+ {%- set first_user_prefix = messages[0]['content'] + '
5
+
6
+ ' -%}
7
+ {%- else -%}
8
+ {%- set first_user_prefix = messages[0]['content'][0]['text'] + '
9
+
10
+ ' -%}
11
+ {%- endif -%}
12
+ {%- set loop_messages = messages[1:] -%}
13
+ {%- else -%}
14
+ {%- set first_user_prefix = "" -%}
15
+ {%- set loop_messages = messages -%}
16
+ {%- endif -%}
17
+ {%- for message in loop_messages -%}
18
+ {%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) -%}
19
+ {{ raise_exception("Conversation roles must alternate user/assistant/user/assistant/...") }}
20
+ {%- endif -%}
21
+ {%- if (message['role'] == 'assistant') -%}
22
+ {%- set role = "model" -%}
23
+ {%- else -%}
24
+ {%- set role = message['role'] -%}
25
+ {%- endif -%}
26
+ {{ '<start_of_turn>' + role + '
27
+ ' + (first_user_prefix if loop.first else "") }}
28
+ {%- if message['content'] is string -%}
29
+ {{ message['content'] | trim }}
30
+ {%- elif message['content'] is iterable -%}
31
+ {%- for item in message['content'] -%}
32
+ {%- if item['type'] == 'image' -%}
33
+ {{ '<start_of_image>' }}
34
+ {%- elif item['type'] == 'text' -%}
35
+ {{ item['text'] | trim }}
36
+ {%- endif -%}
37
+ {%- endfor -%}
38
+ {%- else -%}
39
+ {{ raise_exception("Invalid content type") }}
40
+ {%- endif -%}
41
+ {{ '<end_of_turn>
42
+ ' }}
43
+ {%- endfor -%}
44
+ {%- if add_generation_prompt -%}
45
+ {{'<start_of_turn>model
46
+ '}}
47
+ {%- endif -%}
flare-medgemma-medgemma15-lora/checkpoint-17200/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:184f6a2ec4001c6b02b9ee7055caef68fb0a3cb3b51b1f5dbb04b6f65f037391
3
+ size 79126053
flare-medgemma-medgemma15-lora/checkpoint-17200/processor_config.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "image_processor": {
3
+ "do_convert_rgb": true,
4
+ "do_normalize": true,
5
+ "do_rescale": true,
6
+ "do_resize": true,
7
+ "image_mean": [
8
+ 0.5,
9
+ 0.5,
10
+ 0.5
11
+ ],
12
+ "image_processor_type": "Gemma3ImageProcessor",
13
+ "image_seq_length": 256,
14
+ "image_std": [
15
+ 0.5,
16
+ 0.5,
17
+ 0.5
18
+ ],
19
+ "resample": 2,
20
+ "rescale_factor": 0.00392156862745098,
21
+ "size": {
22
+ "height": 896,
23
+ "width": 896
24
+ }
25
+ },
26
+ "image_seq_length": 256,
27
+ "processor_class": "Gemma3Processor"
28
+ }
flare-medgemma-medgemma15-lora/checkpoint-17200/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:da0937033e6fbb399be54851f544191f001f2927fa7d9280b8deac2d1d9e7e51
3
+ size 14645
flare-medgemma-medgemma15-lora/checkpoint-17200/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b8f4c6e76e7259f0e5399c4131104fb8e4bce6c9f02b1ac528873610791e33c6
3
+ size 1465
flare-medgemma-medgemma15-lora/checkpoint-17200/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:daab2354f8a74e70d70b4d1f804939b68a8c9624dd06cb7858e52dd8970e9726
3
+ size 33384567
flare-medgemma-medgemma15-lora/checkpoint-17200/tokenizer_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "backend": "tokenizers",
3
+ "boi_token": "<start_of_image>",
4
+ "bos_token": "<bos>",
5
+ "clean_up_tokenization_spaces": false,
6
+ "eoi_token": "<end_of_image>",
7
+ "eos_token": "<eos>",
8
+ "image_token": "<image_soft_token>",
9
+ "is_local": false,
10
+ "local_files_only": false,
11
+ "mask_token": "<mask>",
12
+ "model_max_length": 1000000000000000019884624838656,
13
+ "model_specific_special_tokens": {
14
+ "boi_token": "<start_of_image>",
15
+ "eoi_token": "<end_of_image>",
16
+ "image_token": "<image_soft_token>"
17
+ },
18
+ "pad_token": "<pad>",
19
+ "padding_side": "right",
20
+ "processor_class": "Gemma3Processor",
21
+ "sp_model_kwargs": null,
22
+ "spaces_between_special_tokens": false,
23
+ "tokenizer_class": "GemmaTokenizer",
24
+ "unk_token": "<unk>",
25
+ "use_default_system_prompt": false
26
+ }
flare-medgemma-medgemma15-lora/checkpoint-17200/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
flare-medgemma-medgemma15-lora/checkpoint-17200/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e7cafb5cde52b908ce007ba70fb2b52cc4f4dffde9f907f7e5805f932473d809
3
+ size 5905
flare-medgemma-medgemma15-lora/checkpoint-17208/README.md ADDED
@@ -0,0 +1,209 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: google/medgemma-1.5-4b-it
3
+ library_name: peft
4
+ pipeline_tag: text-generation
5
+ tags:
6
+ - base_model:adapter:google/medgemma-1.5-4b-it
7
+ - lora
8
+ - sft
9
+ - transformers
10
+ - trl
11
+ ---
12
+
13
+ # Model Card for Model ID
14
+
15
+ <!-- Provide a quick summary of what the model is/does. -->
16
+
17
+
18
+
19
+ ## Model Details
20
+
21
+ ### Model Description
22
+
23
+ <!-- Provide a longer summary of what this model is. -->
24
+
25
+
26
+
27
+ - **Developed by:** [More Information Needed]
28
+ - **Funded by [optional]:** [More Information Needed]
29
+ - **Shared by [optional]:** [More Information Needed]
30
+ - **Model type:** [More Information Needed]
31
+ - **Language(s) (NLP):** [More Information Needed]
32
+ - **License:** [More Information Needed]
33
+ - **Finetuned from model [optional]:** [More Information Needed]
34
+
35
+ ### Model Sources [optional]
36
+
37
+ <!-- Provide the basic links for the model. -->
38
+
39
+ - **Repository:** [More Information Needed]
40
+ - **Paper [optional]:** [More Information Needed]
41
+ - **Demo [optional]:** [More Information Needed]
42
+
43
+ ## Uses
44
+
45
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
46
+
47
+ ### Direct Use
48
+
49
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
50
+
51
+ [More Information Needed]
52
+
53
+ ### Downstream Use [optional]
54
+
55
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
56
+
57
+ [More Information Needed]
58
+
59
+ ### Out-of-Scope Use
60
+
61
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
62
+
63
+ [More Information Needed]
64
+
65
+ ## Bias, Risks, and Limitations
66
+
67
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
68
+
69
+ [More Information Needed]
70
+
71
+ ### Recommendations
72
+
73
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
74
+
75
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
76
+
77
+ ## How to Get Started with the Model
78
+
79
+ Use the code below to get started with the model.
80
+
81
+ [More Information Needed]
82
+
83
+ ## Training Details
84
+
85
+ ### Training Data
86
+
87
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
88
+
89
+ [More Information Needed]
90
+
91
+ ### Training Procedure
92
+
93
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
94
+
95
+ #### Preprocessing [optional]
96
+
97
+ [More Information Needed]
98
+
99
+
100
+ #### Training Hyperparameters
101
+
102
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
103
+
104
+ #### Speeds, Sizes, Times [optional]
105
+
106
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
107
+
108
+ [More Information Needed]
109
+
110
+ ## Evaluation
111
+
112
+ <!-- This section describes the evaluation protocols and provides the results. -->
113
+
114
+ ### Testing Data, Factors & Metrics
115
+
116
+ #### Testing Data
117
+
118
+ <!-- This should link to a Dataset Card if possible. -->
119
+
120
+ [More Information Needed]
121
+
122
+ #### Factors
123
+
124
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
125
+
126
+ [More Information Needed]
127
+
128
+ #### Metrics
129
+
130
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
131
+
132
+ [More Information Needed]
133
+
134
+ ### Results
135
+
136
+ [More Information Needed]
137
+
138
+ #### Summary
139
+
140
+
141
+
142
+ ## Model Examination [optional]
143
+
144
+ <!-- Relevant interpretability work for the model goes here -->
145
+
146
+ [More Information Needed]
147
+
148
+ ## Environmental Impact
149
+
150
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
151
+
152
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
153
+
154
+ - **Hardware Type:** [More Information Needed]
155
+ - **Hours used:** [More Information Needed]
156
+ - **Cloud Provider:** [More Information Needed]
157
+ - **Compute Region:** [More Information Needed]
158
+ - **Carbon Emitted:** [More Information Needed]
159
+
160
+ ## Technical Specifications [optional]
161
+
162
+ ### Model Architecture and Objective
163
+
164
+ [More Information Needed]
165
+
166
+ ### Compute Infrastructure
167
+
168
+ [More Information Needed]
169
+
170
+ #### Hardware
171
+
172
+ [More Information Needed]
173
+
174
+ #### Software
175
+
176
+ [More Information Needed]
177
+
178
+ ## Citation [optional]
179
+
180
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
181
+
182
+ **BibTeX:**
183
+
184
+ [More Information Needed]
185
+
186
+ **APA:**
187
+
188
+ [More Information Needed]
189
+
190
+ ## Glossary [optional]
191
+
192
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
193
+
194
+ [More Information Needed]
195
+
196
+ ## More Information [optional]
197
+
198
+ [More Information Needed]
199
+
200
+ ## Model Card Authors [optional]
201
+
202
+ [More Information Needed]
203
+
204
+ ## Model Card Contact
205
+
206
+ [More Information Needed]
207
+ ### Framework versions
208
+
209
+ - PEFT 0.19.1
flare-medgemma-medgemma15-lora/checkpoint-17208/adapter_config.json ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alora_invocation_tokens": null,
3
+ "alpha_pattern": {},
4
+ "arrow_config": null,
5
+ "auto_mapping": null,
6
+ "base_model_name_or_path": "google/medgemma-1.5-4b-it",
7
+ "bias": "none",
8
+ "corda_config": null,
9
+ "ensure_weight_tying": false,
10
+ "eva_config": null,
11
+ "exclude_modules": null,
12
+ "fan_in_fan_out": false,
13
+ "inference_mode": true,
14
+ "init_lora_weights": true,
15
+ "layer_replication": null,
16
+ "layers_pattern": null,
17
+ "layers_to_transform": null,
18
+ "loftq_config": {},
19
+ "lora_alpha": 16,
20
+ "lora_bias": false,
21
+ "lora_dropout": 0.05,
22
+ "lora_ga_config": null,
23
+ "megatron_config": null,
24
+ "megatron_core": "megatron.core",
25
+ "modules_to_save": null,
26
+ "peft_type": "LORA",
27
+ "peft_version": "0.19.1",
28
+ "qalora_group_size": 16,
29
+ "r": 16,
30
+ "rank_pattern": {},
31
+ "revision": null,
32
+ "target_modules": [
33
+ "gate_proj",
34
+ "down_proj",
35
+ "fc2",
36
+ "k_proj",
37
+ "fc1",
38
+ "out_proj",
39
+ "v_proj",
40
+ "up_proj",
41
+ "o_proj",
42
+ "q_proj"
43
+ ],
44
+ "target_parameters": null,
45
+ "task_type": "CAUSAL_LM",
46
+ "trainable_token_indices": null,
47
+ "use_bdlora": null,
48
+ "use_dora": false,
49
+ "use_qalora": false,
50
+ "use_rslora": false
51
+ }
flare-medgemma-medgemma15-lora/checkpoint-17208/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b2ef5d91bdfc89fa724c45ab3248ac37dc7baaf3b80b87107527b5d739fed28d
3
+ size 77116432
flare-medgemma-medgemma15-lora/checkpoint-17208/chat_template.jinja ADDED
@@ -0,0 +1,47 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {{ bos_token }}
2
+ {%- if messages[0]['role'] == 'system' -%}
3
+ {%- if messages[0]['content'] is string -%}
4
+ {%- set first_user_prefix = messages[0]['content'] + '
5
+
6
+ ' -%}
7
+ {%- else -%}
8
+ {%- set first_user_prefix = messages[0]['content'][0]['text'] + '
9
+
10
+ ' -%}
11
+ {%- endif -%}
12
+ {%- set loop_messages = messages[1:] -%}
13
+ {%- else -%}
14
+ {%- set first_user_prefix = "" -%}
15
+ {%- set loop_messages = messages -%}
16
+ {%- endif -%}
17
+ {%- for message in loop_messages -%}
18
+ {%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) -%}
19
+ {{ raise_exception("Conversation roles must alternate user/assistant/user/assistant/...") }}
20
+ {%- endif -%}
21
+ {%- if (message['role'] == 'assistant') -%}
22
+ {%- set role = "model" -%}
23
+ {%- else -%}
24
+ {%- set role = message['role'] -%}
25
+ {%- endif -%}
26
+ {{ '<start_of_turn>' + role + '
27
+ ' + (first_user_prefix if loop.first else "") }}
28
+ {%- if message['content'] is string -%}
29
+ {{ message['content'] | trim }}
30
+ {%- elif message['content'] is iterable -%}
31
+ {%- for item in message['content'] -%}
32
+ {%- if item['type'] == 'image' -%}
33
+ {{ '<start_of_image>' }}
34
+ {%- elif item['type'] == 'text' -%}
35
+ {{ item['text'] | trim }}
36
+ {%- endif -%}
37
+ {%- endfor -%}
38
+ {%- else -%}
39
+ {{ raise_exception("Invalid content type") }}
40
+ {%- endif -%}
41
+ {{ '<end_of_turn>
42
+ ' }}
43
+ {%- endfor -%}
44
+ {%- if add_generation_prompt -%}
45
+ {{'<start_of_turn>model
46
+ '}}
47
+ {%- endif -%}
flare-medgemma-medgemma15-lora/checkpoint-17208/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:31a742e004fc9dac896c9f49e2edceb89070d779d5e22e67c66a85efec22e3b5
3
+ size 79126053
flare-medgemma-medgemma15-lora/checkpoint-17208/processor_config.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "image_processor": {
3
+ "do_convert_rgb": true,
4
+ "do_normalize": true,
5
+ "do_rescale": true,
6
+ "do_resize": true,
7
+ "image_mean": [
8
+ 0.5,
9
+ 0.5,
10
+ 0.5
11
+ ],
12
+ "image_processor_type": "Gemma3ImageProcessor",
13
+ "image_seq_length": 256,
14
+ "image_std": [
15
+ 0.5,
16
+ 0.5,
17
+ 0.5
18
+ ],
19
+ "resample": 2,
20
+ "rescale_factor": 0.00392156862745098,
21
+ "size": {
22
+ "height": 896,
23
+ "width": 896
24
+ }
25
+ },
26
+ "image_seq_length": 256,
27
+ "processor_class": "Gemma3Processor"
28
+ }
flare-medgemma-medgemma15-lora/checkpoint-17208/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:98e4b06e9cd2966f055522ac64831c7af7534d73ac7d3362714cecabc7f0130f
3
+ size 14645
flare-medgemma-medgemma15-lora/checkpoint-17208/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a23a73e235336b4b6f2397db719a95907cb035346f18c26391475acde4fafb26
3
+ size 1465
flare-medgemma-medgemma15-lora/checkpoint-17208/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:daab2354f8a74e70d70b4d1f804939b68a8c9624dd06cb7858e52dd8970e9726
3
+ size 33384567
flare-medgemma-medgemma15-lora/checkpoint-17208/tokenizer_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "backend": "tokenizers",
3
+ "boi_token": "<start_of_image>",
4
+ "bos_token": "<bos>",
5
+ "clean_up_tokenization_spaces": false,
6
+ "eoi_token": "<end_of_image>",
7
+ "eos_token": "<eos>",
8
+ "image_token": "<image_soft_token>",
9
+ "is_local": false,
10
+ "local_files_only": false,
11
+ "mask_token": "<mask>",
12
+ "model_max_length": 1000000000000000019884624838656,
13
+ "model_specific_special_tokens": {
14
+ "boi_token": "<start_of_image>",
15
+ "eoi_token": "<end_of_image>",
16
+ "image_token": "<image_soft_token>"
17
+ },
18
+ "pad_token": "<pad>",
19
+ "padding_side": "right",
20
+ "processor_class": "Gemma3Processor",
21
+ "sp_model_kwargs": null,
22
+ "spaces_between_special_tokens": false,
23
+ "tokenizer_class": "GemmaTokenizer",
24
+ "unk_token": "<unk>",
25
+ "use_default_system_prompt": false
26
+ }
flare-medgemma-medgemma15-lora/checkpoint-17208/trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
flare-medgemma-medgemma15-lora/checkpoint-17208/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e7cafb5cde52b908ce007ba70fb2b52cc4f4dffde9f907f7e5805f932473d809
3
+ size 5905