Hugging Face's logo

PaulD
/
llama3_false_positives_0609_KTO_hp_screening_seeds

PEFT
Safetensors
trl
kto
Generated from Trainer
Model card Files Community
PaulD commited on
Commit
704b295
•
1 Parent(s): 73ad605

End of training

Browse files
Files changed (13) hide show
  1. README.md +14 -14
  2. adapter_config.json +1 -1
  3. adapter_model.safetensors +1 -1
  4. metrics.jsonl +4 -0
  5. metrics_epoch_0.9411764705882353_fold_0_lr_1e-05_seed_9012.json +1 -0
  6. metrics_epoch_1.9607843137254903_fold_0_lr_1e-05_seed_9012.json +1 -0
  7. metrics_epoch_2.980392156862745_fold_0_lr_1e-05_seed_9012.json +1 -0
  8. metrics_epoch_3.764705882352941_fold_0_lr_1e-05_seed_9012.json +1 -0
  9. results_epoch_0.9411764705882353_fold_0_lr_1e-05_seed_9012.json +0 -0
  10. results_epoch_1.9607843137254903_fold_0_lr_1e-05_seed_9012.json +0 -0
  11. results_epoch_2.980392156862745_fold_0_lr_1e-05_seed_9012.json +0 -0
  12. results_epoch_3.764705882352941_fold_0_lr_1e-05_seed_9012.json +0 -0
  13. training_args.bin +1 -1
README.md CHANGED
@@ -18,13 +18,13 @@ should probably proofread and complete it, then remove this comment. -->
18
 
19
  This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on the None dataset.
20
  It achieves the following results on the evaluation set:
21
- - Loss: 0.7043
22
- - Eval/rewards/chosen: 2.7139
23
- - Eval/logps/chosen: -180.4410
24
- - Eval/rewards/rejected: 2.5251
25
- - Eval/logps/rejected: -196.3258
26
- - Eval/rewards/margins: 0.1888
27
- - Eval/kl: 22.1345
28
 
29
  ## Model description
30
 
@@ -43,7 +43,7 @@ More information needed
43
  ### Training hyperparameters
44
 
45
  The following hyperparameters were used during training:
46
- - learning_rate: 0.0001
47
  - train_batch_size: 1
48
  - eval_batch_size: 2
49
  - seed: 9012
@@ -56,12 +56,12 @@ The following hyperparameters were used during training:
56
 
57
  ### Training results
58
 
59
- | Training Loss | Epoch | Step | Validation Loss | |
60
- |:-------------:|:------:|:----:|:---------------:|:-------:|
61
- | 0.5502 | 0.9412 | 12 | 0.6760 | 8.5385 |
62
- | 0.3996 | 1.9608 | 25 | 0.6791 | 17.9477 |
63
- | 0.2636 | 2.9804 | 38 | 0.7076 | 21.9801 |
64
- | 0.34 | 3.7647 | 48 | 0.7043 | 22.1345 |
65
 
66
 
67
  ### Framework versions
 
18
 
19
  This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on the None dataset.
20
  It achieves the following results on the evaluation set:
21
+ - Loss: 0.7271
22
+ - Eval/rewards/chosen: 0.0966
23
+ - Eval/logps/chosen: -206.6140
24
+ - Eval/rewards/rejected: 0.0752
25
+ - Eval/logps/rejected: -220.8250
26
+ - Eval/rewards/margins: 0.0214
27
+ - Eval/kl: 0.8580
28
 
29
  ## Model description
30
 
 
43
  ### Training hyperparameters
44
 
45
  The following hyperparameters were used during training:
46
+ - learning_rate: 1e-05
47
  - train_batch_size: 1
48
  - eval_batch_size: 2
49
  - seed: 9012
 
56
 
57
  ### Training results
58
 
59
+ | Training Loss | Epoch | Step | Validation Loss | |
60
+ |:-------------:|:------:|:----:|:---------------:|:------:|
61
+ | 0.5994 | 0.9412 | 12 | 0.7301 | 0.1561 |
62
+ | 0.6 | 1.9608 | 25 | 0.7256 | 0.6149 |
63
+ | 0.5831 | 2.9804 | 38 | 0.7234 | 0.8062 |
64
+ | 0.6595 | 3.7647 | 48 | 0.7271 | 0.8580 |
65
 
66
 
67
  ### Framework versions
adapter_config.json CHANGED
@@ -20,9 +20,9 @@
20
  "rank_pattern": {},
21
  "revision": null,
22
  "target_modules": [
23
- "q_proj",
24
  "v_proj",
25
  "k_proj",
 
26
  "o_proj"
27
  ],
28
  "task_type": "CAUSAL_LM",
 
20
  "rank_pattern": {},
21
  "revision": null,
22
  "target_modules": [
 
23
  "v_proj",
24
  "k_proj",
25
+ "q_proj",
26
  "o_proj"
27
  ],
28
  "task_type": "CAUSAL_LM",
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d6a162993557bf76220bd80f53f3d490e22e78e084f5148682e238148f96cf40
3
  size 27297544
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:46fa99059c311a3f0ce1324a5f094226885b50b034ddaa30cc8259a753fb5d1b
3
  size 27297544
metrics.jsonl CHANGED
@@ -18,3 +18,7 @@
18
  {"epoch": 1.9607843137254903, "precision": 0.6666666646464646, "recall": 0.9166666628472222, "fold": 0}
19
  {"epoch": 2.980392156862745, "precision": 0.6451612882414152, "recall": 0.8333333298611111, "fold": 0}
20
  {"epoch": 3.764705882352941, "precision": 0.6451612882414152, "recall": 0.8333333298611111, "fold": 0}
 
 
 
 
 
18
  {"epoch": 1.9607843137254903, "precision": 0.6666666646464646, "recall": 0.9166666628472222, "fold": 0}
19
  {"epoch": 2.980392156862745, "precision": 0.6451612882414152, "recall": 0.8333333298611111, "fold": 0}
20
  {"epoch": 3.764705882352941, "precision": 0.6451612882414152, "recall": 0.8333333298611111, "fold": 0}
21
+ {"epoch": 0.9411764705882353, "precision": 0.4313725481737793, "recall": 0.9166666628472222, "fold": 0}
22
+ {"epoch": 1.9607843137254903, "precision": 0.4999999989130435, "recall": 0.9583333293402777, "fold": 0}
23
+ {"epoch": 2.980392156862745, "precision": 0.4693877541441066, "recall": 0.9583333293402777, "fold": 0}
24
+ {"epoch": 3.764705882352941, "precision": 0.4423076914571006, "recall": 0.9583333293402777, "fold": 0}
metrics_epoch_0.9411764705882353_fold_0_lr_1e-05_seed_9012.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"epoch": 0.9411764705882353, "precision": 0.4313725481737793, "recall": 0.9166666628472222, "fold": 0}
metrics_epoch_1.9607843137254903_fold_0_lr_1e-05_seed_9012.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"epoch": 1.9607843137254903, "precision": 0.4999999989130435, "recall": 0.9583333293402777, "fold": 0}
metrics_epoch_2.980392156862745_fold_0_lr_1e-05_seed_9012.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"epoch": 2.980392156862745, "precision": 0.4693877541441066, "recall": 0.9583333293402777, "fold": 0}
metrics_epoch_3.764705882352941_fold_0_lr_1e-05_seed_9012.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"epoch": 3.764705882352941, "precision": 0.4423076914571006, "recall": 0.9583333293402777, "fold": 0}
results_epoch_0.9411764705882353_fold_0_lr_1e-05_seed_9012.json ADDED
The diff for this file is too large to render. See raw diff
 
results_epoch_1.9607843137254903_fold_0_lr_1e-05_seed_9012.json ADDED
The diff for this file is too large to render. See raw diff
 
results_epoch_2.980392156862745_fold_0_lr_1e-05_seed_9012.json ADDED
The diff for this file is too large to render. See raw diff
 
results_epoch_3.764705882352941_fold_0_lr_1e-05_seed_9012.json ADDED
The diff for this file is too large to render. See raw diff
 
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9f83b5814ea47fa9e8f412336fab0b6a313875271cb82b9c6fe937db904a888d
3
  size 5688
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c5587e3a2b9c35943509b928a682c25266fdb7918b7311a7a43bc4b34da1398a
3
  size 5688