Hugging Face's logo

PaulD
/
llama3_false_positives_1010_KTO_hp_screening_seeds

PEFT
Safetensors
trl
kto
Generated from Trainer
Model card Files Community
PaulD commited on
Commit
7db5dcf
·
verified ·
1 Parent(s): b331263

End of training

Browse files
Files changed (9) hide show
  1. README.md +10 -10
  2. adapter_config.json +1 -1
  3. adapter_model.safetensors +1 -1
  4. metrics.jsonl +2 -0
  5. metrics_epoch_0.96_fold_0_lr_0.0001_seed_1234_weight_2.0.json +1 -0
  6. metrics_epoch_2.0_fold_0_lr_0.0001_seed_1234_weight_2.0.json +1 -0
  7. results_epoch_0.96_fold_0_lr_0.0001_seed_1234_weight_2.0.json +0 -0
  8. results_epoch_2.0_fold_0_lr_0.0001_seed_1234_weight_2.0.json +0 -0
  9. 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: 0.6955
22
- - Eval/rewards/chosen: 2.6328
23
- - Eval/logps/chosen: -171.1206
24
- - Eval/rewards/rejected: 3.7940
25
- - Eval/logps/rejected: -216.4652
26
- - Eval/rewards/margins: -1.1613
27
- - Eval/kl: 30.4319
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: 0.001
47
  - train_batch_size: 1
48
  - eval_batch_size: 2
49
  - seed: 1234
@@ -58,8 +58,8 @@ The following hyperparameters were used during training:
58
 
59
  | Training Loss | Epoch | Step | Validation Loss | |
60
  |:-------------:|:-----:|:----:|:---------------:|:-------:|
61
- | 0.2388 | 0.96 | 12 | 0.7812 | 18.6117 |
62
- | 0.2319 | 2.0 | 25 | 0.6955 | 30.4319 |
63
 
64
 
65
  ### 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: 0.6324
22
+ - Eval/rewards/chosen: 1.4100
23
+ - Eval/logps/chosen: -183.3489
24
+ - Eval/rewards/rejected: 1.3516
25
+ - Eval/logps/rejected: -240.8897
26
+ - Eval/rewards/margins: 0.0584
27
+ - Eval/kl: 14.3395
28
 
29
  ## Model description
30
 
 
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: 1234
 
58
 
59
  | Training Loss | Epoch | Step | Validation Loss | |
60
  |:-------------:|:-----:|:----:|:---------------:|:-------:|
61
+ | 0.6561 | 0.96 | 12 | 0.6281 | 0.1571 |
62
+ | 0.3245 | 2.0 | 25 | 0.6324 | 14.3395 |
63
 
64
 
65
  ### Framework versions
adapter_config.json CHANGED
@@ -21,8 +21,8 @@
21
  "revision": null,
22
  "target_modules": [
23
  "q_proj",
24
- "k_proj",
25
  "v_proj",
 
26
  "o_proj"
27
  ],
28
  "task_type": "CAUSAL_LM",
 
21
  "revision": null,
22
  "target_modules": [
23
  "q_proj",
 
24
  "v_proj",
25
+ "k_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 sha256:17ac95fd247bbc9c5961b8858bb287bda1ecd7edd47316ff3c35b13d78a9704d
3
  size 27297544
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c19d7e7815d0a7ac4e697c4b94d61f128a09b7bda2c1b68497d125ee7f49616f
3
  size 27297544
metrics.jsonl CHANGED
@@ -1,2 +1,4 @@
1
  {"epoch": 0.96, "precision": 0.0, "recall": 0.0, "fold": 0}
2
  {"epoch": 2.0, "precision": 0.0, "recall": 0.0, "fold": 0}
 
 
 
1
  {"epoch": 0.96, "precision": 0.0, "recall": 0.0, "fold": 0}
2
  {"epoch": 2.0, "precision": 0.0, "recall": 0.0, "fold": 0}
3
+ {"epoch": 0.96, "precision": 0.6363636349173554, "recall": 0.9333333302222222, "fold": 0}
4
+ {"epoch": 2.0, "precision": 0.8529411739619377, "recall": 0.9666666634444444, "fold": 0}
metrics_epoch_0.96_fold_0_lr_0.0001_seed_1234_weight_2.0.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"epoch": 0.96, "precision": 0.6363636349173554, "recall": 0.9333333302222222, "fold": 0}
metrics_epoch_2.0_fold_0_lr_0.0001_seed_1234_weight_2.0.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"epoch": 2.0, "precision": 0.8529411739619377, "recall": 0.9666666634444444, "fold": 0}
results_epoch_0.96_fold_0_lr_0.0001_seed_1234_weight_2.0.json ADDED
The diff for this file is too large to render. See raw diff
 
results_epoch_2.0_fold_0_lr_0.0001_seed_1234_weight_2.0.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 sha256:6734f2b2355882795feb35e63f250abe4f7a4ff8700e1545151016680fa19064
3
  size 5688
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3b8a74797043c53ab902975cf24619e88f2c8d298a72217fb300e0b4c8618d38
3
  size 5688