Hugging Face's logo

tsavage68
/
MedQA_L3_1000steps_1e5rate_05beta_CSFTDPO

Text Generation
Transformers
Safetensors
llama
trl
dpo
Generated from Trainer
conversational
text-generation-inference
Inference Endpoints
Model card Files Community
tsavage68 commited on
Commit
9cdfb43
•
1 Parent(s): 335995e

End of training

Browse files
Files changed (15) hide show
  1. README.md +34 -34
  2. config.json +2 -2
  3. final_checkpoint/config.json +2 -2
  4. final_checkpoint/generation_config.json +1 -1
  5. final_checkpoint/model-00001-of-00004.safetensors +1 -1
  6. final_checkpoint/model-00002-of-00004.safetensors +1 -1
  7. final_checkpoint/model-00003-of-00004.safetensors +1 -1
  8. final_checkpoint/model-00004-of-00004.safetensors +1 -1
  9. generation_config.json +1 -1
  10. model-00001-of-00004.safetensors +1 -1
  11. model-00002-of-00004.safetensors +1 -1
  12. model-00003-of-00004.safetensors +1 -1
  13. model-00004-of-00004.safetensors +1 -1
  14. tokenizer_config.json +5 -1
  15. training_args.bin +1 -1
README.md CHANGED
@@ -1,31 +1,31 @@
1
  ---
2
  license: llama3
3
- base_model: meta-llama/Meta-Llama-3-8B-Instruct
4
  tags:
5
  - trl
6
  - dpo
7
  - generated_from_trainer
8
  model-index:
9
- - name: MedQA_L3_1000steps_1e5rate_03beta_CSFTDPO
10
  results: []
11
  ---
12
 
13
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
14
  should probably proofread and complete it, then remove this comment. -->
15
 
16
- # MedQA_L3_1000steps_1e5rate_03beta_CSFTDPO
17
 
18
- 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 an unknown dataset.
19
  It achieves the following results on the evaluation set:
20
- - Loss: 3.1515
21
- - Rewards/chosen: -14.8095
22
- - Rewards/rejected: -13.6020
23
- - Rewards/accuracies: 0.4176
24
- - Rewards/margins: -1.2075
25
- - Logps/rejected: -48.5205
26
- - Logps/chosen: -47.8417
27
- - Logits/rejected: -0.3920
28
- - Logits/chosen: -0.3920
29
 
30
  ## Model description
31
 
@@ -59,31 +59,31 @@ The following hyperparameters were used during training:
59
 
60
  | Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
61
  |:-------------:|:------:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
62
- | 1.1843 | 0.0489 | 50 | 1.4952 | -9.4284 | -9.7251 | 0.5604 | 0.2967 | -40.7667 | -37.0794 | -0.9884 | -0.9876 |
63
- | 2.9825 | 0.0977 | 100 | 2.9340 | -11.3951 | -10.1604 | 0.3956 | -1.2347 | -41.6373 | -41.0129 | -0.7061 | -0.7063 |
64
- | 5.07 | 0.1466 | 150 | 5.8445 | -33.3133 | -32.2334 | 0.4330 | -1.0798 | -85.7834 | -84.8492 | 0.6418 | 0.6418 |
65
- | 3.8509 | 0.1954 | 200 | 4.3552 | -19.5431 | -18.0372 | 0.3868 | -1.5059 | -57.3910 | -57.3089 | 0.4504 | 0.4502 |
66
- | 2.4016 | 0.2443 | 250 | 4.1219 | -19.6576 | -18.2218 | 0.4154 | -1.4358 | -57.7601 | -57.5379 | -0.6569 | -0.6572 |
67
- | 6.0228 | 0.2931 | 300 | 3.6535 | -22.0862 | -20.5713 | 0.4044 | -1.5149 | -62.4591 | -62.3951 | -0.5981 | -0.5981 |
68
- | 6.0198 | 0.3420 | 350 | 4.5591 | -27.2649 | -25.8931 | 0.3956 | -1.3718 | -73.1028 | -72.7525 | -0.4660 | -0.4661 |
69
- | 4.3784 | 0.3908 | 400 | 3.8339 | -16.6518 | -14.9972 | 0.4044 | -1.6546 | -51.3108 | -51.5262 | -0.5637 | -0.5639 |
70
- | 4.9465 | 0.4397 | 450 | 3.5058 | -18.5653 | -17.0632 | 0.3912 | -1.5021 | -55.4430 | -55.3533 | -0.8423 | -0.8424 |
71
- | 5.0702 | 0.4885 | 500 | 3.7225 | -18.6932 | -17.3825 | 0.4022 | -1.3108 | -56.0814 | -55.6091 | -1.0248 | -1.0249 |
72
- | 3.3013 | 0.5374 | 550 | 3.5310 | -17.9328 | -16.5937 | 0.4242 | -1.3391 | -54.5040 | -54.0883 | -0.4439 | -0.4440 |
73
- | 3.6542 | 0.5862 | 600 | 3.6084 | -16.4171 | -15.1231 | 0.4110 | -1.2939 | -51.5628 | -51.0568 | -0.3546 | -0.3546 |
74
- | 3.6485 | 0.6351 | 650 | 3.5138 | -18.0930 | -16.6063 | 0.3890 | -1.4867 | -54.5292 | -54.4087 | -0.6487 | -0.6487 |
75
- | 3.2954 | 0.6839 | 700 | 3.6009 | -15.0224 | -13.6768 | 0.4110 | -1.3456 | -48.6701 | -48.2675 | -0.3265 | -0.3266 |
76
- | 3.5098 | 0.7328 | 750 | 3.3569 | -16.2005 | -14.8261 | 0.4220 | -1.3744 | -50.9687 | -50.6236 | -0.4849 | -0.4850 |
77
- | 2.6183 | 0.7816 | 800 | 3.1964 | -14.9792 | -13.7545 | 0.4110 | -1.2247 | -48.8255 | -48.1811 | -0.4045 | -0.4045 |
78
- | 3.4615 | 0.8305 | 850 | 3.1868 | -14.5397 | -13.3293 | 0.4176 | -1.2105 | -47.9750 | -47.3021 | -0.3691 | -0.3691 |
79
- | 2.4743 | 0.8793 | 900 | 3.1532 | -14.7973 | -13.5910 | 0.4132 | -1.2063 | -48.4985 | -47.8173 | -0.3918 | -0.3918 |
80
- | 2.337 | 0.9282 | 950 | 3.1514 | -14.8091 | -13.6016 | 0.4176 | -1.2075 | -48.5197 | -47.8408 | -0.3921 | -0.3920 |
81
- | 2.2422 | 0.9770 | 1000 | 3.1515 | -14.8095 | -13.6020 | 0.4176 | -1.2075 | -48.5205 | -47.8417 | -0.3920 | -0.3920 |
82
 
83
 
84
  ### Framework versions
85
 
86
- - Transformers 4.41.0
87
  - Pytorch 2.0.0+cu117
88
  - Datasets 2.19.1
89
  - Tokenizers 0.19.1
 
1
  ---
2
  license: llama3
3
+ base_model: tsavage68/MedQA_L3_1000steps_1e6rate_SFT
4
  tags:
5
  - trl
6
  - dpo
7
  - generated_from_trainer
8
  model-index:
9
+ - name: MedQA_L3_1000steps_1e5rate_05beta_CSFTDPO
10
  results: []
11
  ---
12
 
13
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
14
  should probably proofread and complete it, then remove this comment. -->
15
 
16
+ # MedQA_L3_1000steps_1e5rate_05beta_CSFTDPO
17
 
18
+ This model is a fine-tuned version of [tsavage68/MedQA_L3_1000steps_1e6rate_SFT](https://huggingface.co/tsavage68/MedQA_L3_1000steps_1e6rate_SFT) on an unknown dataset.
19
  It achieves the following results on the evaluation set:
20
+ - Loss: 2.7867
21
+ - Rewards/chosen: -10.2874
22
+ - Rewards/rejected: -9.4675
23
+ - Rewards/accuracies: 0.4330
24
+ - Rewards/margins: -0.8198
25
+ - Logps/rejected: -52.7899
26
+ - Logps/chosen: -51.9033
27
+ - Logits/rejected: -0.3129
28
+ - Logits/chosen: -0.3128
29
 
30
  ## Model description
31
 
 
59
 
60
  | Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
61
  |:-------------:|:------:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
62
+ | 0.9373 | 0.0489 | 50 | 1.5325 | 0.6891 | -0.1945 | 0.5912 | 0.8836 | -34.2439 | -29.9504 | -1.1200 | -1.1197 |
63
+ | 3.7169 | 0.0977 | 100 | 3.7845 | -9.7504 | -8.8431 | 0.4527 | -0.9074 | -51.5409 | -50.8294 | -0.6137 | -0.6138 |
64
+ | 5.2014 | 0.1466 | 150 | 5.2600 | -22.3993 | -21.8605 | 0.4681 | -0.5389 | -77.5758 | -76.1272 | -1.3215 | -1.3217 |
65
+ | 5.4743 | 0.1954 | 200 | 3.9034 | -7.1491 | -6.2277 | 0.4176 | -0.9214 | -46.3103 | -45.6268 | -0.6483 | -0.6486 |
66
+ | 3.0731 | 0.2443 | 250 | 4.1865 | -11.6364 | -10.1791 | 0.4198 | -1.4572 | -54.2131 | -54.6012 | -0.7051 | -0.7056 |
67
+ | 5.7952 | 0.2931 | 300 | 3.6683 | -9.2381 | -7.9895 | 0.4264 | -1.2486 | -49.8338 | -49.8046 | -0.4055 | -0.4058 |
68
+ | 3.8474 | 0.3420 | 350 | 3.4898 | -12.7687 | -11.9414 | 0.4132 | -0.8274 | -57.7376 | -56.8660 | -0.8625 | -0.8625 |
69
+ | 5.5721 | 0.3908 | 400 | 3.4194 | -13.5468 | -12.3658 | 0.4044 | -1.1810 | -58.5864 | -58.4221 | -0.8921 | -0.8922 |
70
+ | 6.0929 | 0.4397 | 450 | 3.4518 | -12.5599 | -11.2787 | 0.4132 | -1.2812 | -56.4122 | -56.4483 | -0.6596 | -0.6596 |
71
+ | 5.4036 | 0.4885 | 500 | 3.4349 | -13.3250 | -12.3700 | 0.4264 | -0.9550 | -58.5948 | -57.9785 | -0.4398 | -0.4397 |
72
+ | 4.2614 | 0.5374 | 550 | 3.4447 | -13.2741 | -12.0523 | 0.4132 | -1.2218 | -57.9595 | -57.8767 | -0.2318 | -0.2318 |
73
+ | 5.0683 | 0.5862 | 600 | 3.6325 | -10.9169 | -9.7136 | 0.4242 | -1.2033 | -53.2821 | -53.1624 | 0.0024 | 0.0023 |
74
+ | 2.8041 | 0.6351 | 650 | 3.3753 | -13.7510 | -12.4756 | 0.4110 | -1.2754 | -58.8060 | -58.8306 | -0.4253 | -0.4254 |
75
+ | 2.852 | 0.6839 | 700 | 3.2123 | -11.3782 | -10.1837 | 0.4132 | -1.1945 | -54.2221 | -54.0849 | -0.3353 | -0.3353 |
76
+ | 3.1506 | 0.7328 | 750 | 2.9861 | -10.9246 | -9.9019 | 0.4198 | -1.0227 | -53.6587 | -53.1778 | -0.3577 | -0.3577 |
77
+ | 2.9206 | 0.7816 | 800 | 2.8476 | -10.3118 | -9.4465 | 0.4264 | -0.8653 | -52.7479 | -51.9522 | -0.2881 | -0.2880 |
78
+ | 3.6047 | 0.8305 | 850 | 2.8115 | -10.1979 | -9.3565 | 0.4308 | -0.8414 | -52.5679 | -51.7243 | -0.3016 | -0.3015 |
79
+ | 2.4799 | 0.8793 | 900 | 2.7874 | -10.3005 | -9.4828 | 0.4308 | -0.8177 | -52.8204 | -51.9295 | -0.3147 | -0.3146 |
80
+ | 2.8467 | 0.9282 | 950 | 2.7864 | -10.2878 | -9.4711 | 0.4330 | -0.8167 | -52.7969 | -51.9040 | -0.3132 | -0.3130 |
81
+ | 2.2638 | 0.9770 | 1000 | 2.7867 | -10.2874 | -9.4675 | 0.4330 | -0.8198 | -52.7899 | -51.9033 | -0.3129 | -0.3128 |
82
 
83
 
84
  ### Framework versions
85
 
86
+ - Transformers 4.41.1
87
  - Pytorch 2.0.0+cu117
88
  - Datasets 2.19.1
89
  - Tokenizers 0.19.1
config.json CHANGED
@@ -1,5 +1,5 @@
1
  {
2
- "_name_or_path": "meta-llama/Meta-Llama-3-8B-Instruct",
3
  "architectures": [
4
  "LlamaForCausalLM"
5
  ],
@@ -23,7 +23,7 @@
23
  "rope_theta": 500000.0,
24
  "tie_word_embeddings": false,
25
  "torch_dtype": "float16",
26
- "transformers_version": "4.41.0",
27
  "use_cache": false,
28
  "vocab_size": 128256
29
  }
 
1
  {
2
+ "_name_or_path": "tsavage68/MedQA_L3_1000steps_1e6rate_SFT",
3
  "architectures": [
4
  "LlamaForCausalLM"
5
  ],
 
23
  "rope_theta": 500000.0,
24
  "tie_word_embeddings": false,
25
  "torch_dtype": "float16",
26
+ "transformers_version": "4.41.1",
27
  "use_cache": false,
28
  "vocab_size": 128256
29
  }
final_checkpoint/config.json CHANGED
@@ -1,5 +1,5 @@
1
  {
2
- "_name_or_path": "meta-llama/Meta-Llama-3-8B-Instruct",
3
  "architectures": [
4
  "LlamaForCausalLM"
5
  ],
@@ -23,7 +23,7 @@
23
  "rope_theta": 500000.0,
24
  "tie_word_embeddings": false,
25
  "torch_dtype": "float16",
26
- "transformers_version": "4.41.0",
27
  "use_cache": false,
28
  "vocab_size": 128256
29
  }
 
1
  {
2
+ "_name_or_path": "tsavage68/MedQA_L3_1000steps_1e6rate_SFT",
3
  "architectures": [
4
  "LlamaForCausalLM"
5
  ],
 
23
  "rope_theta": 500000.0,
24
  "tie_word_embeddings": false,
25
  "torch_dtype": "float16",
26
+ "transformers_version": "4.41.1",
27
  "use_cache": false,
28
  "vocab_size": 128256
29
  }
final_checkpoint/generation_config.json CHANGED
@@ -8,5 +8,5 @@
8
  "max_length": 4096,
9
  "temperature": 0.6,
10
  "top_p": 0.9,
11
- "transformers_version": "4.41.0"
12
  }
 
8
  "max_length": 4096,
9
  "temperature": 0.6,
10
  "top_p": 0.9,
11
+ "transformers_version": "4.41.1"
12
  }
final_checkpoint/model-00001-of-00004.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:8f84458a09371132b21dc39246c63a6fb19268b72cb9695a188e08f306624772
3
  size 4976698592
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e3bfe0d189b8c25d74449c7cde2e7b6ff9d0b345bf0acb6eab5646a5b95eb3a8
3
  size 4976698592
final_checkpoint/model-00002-of-00004.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9dcb5e0daa0cdc4cf8274791b8f56ac1ae9cbed17e26f5e9b44fbae5ecc20dcb
3
  size 4999802616
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:68185401033d9802234331da2b510aff2dd191e72c29b677791f018bab421c1d
3
  size 4999802616
final_checkpoint/model-00003-of-00004.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5c39383df2cf91192331c7b99ab2068e46494ad5c1ac209be9daf92e129ca72c
3
  size 4915916080
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c20b5851739334eaad060d01a9ee28f037918f27a8f6e9189d3465c893b74955
3
  size 4915916080
final_checkpoint/model-00004-of-00004.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e1ea0d908c8e866b37b6849c93d82a80d564f9917fe2c91b1ba5ce097ed15598
3
  size 1168138808
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:13741957043d65eff98b808945ae5939e51dad50151dcf60b7f608530d9db3d3
3
  size 1168138808
generation_config.json CHANGED
@@ -8,5 +8,5 @@
8
  "max_length": 4096,
9
  "temperature": 0.6,
10
  "top_p": 0.9,
11
- "transformers_version": "4.41.0"
12
  }
 
8
  "max_length": 4096,
9
  "temperature": 0.6,
10
  "top_p": 0.9,
11
+ "transformers_version": "4.41.1"
12
  }
model-00001-of-00004.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:8f84458a09371132b21dc39246c63a6fb19268b72cb9695a188e08f306624772
3
  size 4976698592
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e3bfe0d189b8c25d74449c7cde2e7b6ff9d0b345bf0acb6eab5646a5b95eb3a8
3
  size 4976698592
model-00002-of-00004.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9dcb5e0daa0cdc4cf8274791b8f56ac1ae9cbed17e26f5e9b44fbae5ecc20dcb
3
  size 4999802616
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:68185401033d9802234331da2b510aff2dd191e72c29b677791f018bab421c1d
3
  size 4999802616
model-00003-of-00004.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5c39383df2cf91192331c7b99ab2068e46494ad5c1ac209be9daf92e129ca72c
3
  size 4915916080
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c20b5851739334eaad060d01a9ee28f037918f27a8f6e9189d3465c893b74955
3
  size 4915916080
model-00004-of-00004.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e1ea0d908c8e866b37b6849c93d82a80d564f9917fe2c91b1ba5ce097ed15598
3
  size 1168138808
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:13741957043d65eff98b808945ae5939e51dad50151dcf60b7f608530d9db3d3
3
  size 1168138808
tokenizer_config.json CHANGED
@@ -2053,11 +2053,15 @@
2053
  "chat_template": "{% set loop_messages = messages %}{% for message in loop_messages %}{% set content = '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' %}{% if loop.index0 == 0 %}{% set content = bos_token + content %}{% endif %}{{ content }}{% endfor %}{% if add_generation_prompt %}{{ '<|start_header_id|>assistant<|end_header_id|>\n\n' }}{% endif %}",
2054
  "clean_up_tokenization_spaces": true,
2055
  "eos_token": "<|eot_id|>",
 
2056
  "model_input_names": [
2057
  "input_ids",
2058
  "attention_mask"
2059
  ],
2060
  "model_max_length": 1000000000000000019884624838656,
2061
  "pad_token": "<|eot_id|>",
2062
- "tokenizer_class": "PreTrainedTokenizerFast"
 
 
 
2063
  }
 
2053
  "chat_template": "{% set loop_messages = messages %}{% for message in loop_messages %}{% set content = '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' %}{% if loop.index0 == 0 %}{% set content = bos_token + content %}{% endif %}{{ content }}{% endfor %}{% if add_generation_prompt %}{{ '<|start_header_id|>assistant<|end_header_id|>\n\n' }}{% endif %}",
2054
  "clean_up_tokenization_spaces": true,
2055
  "eos_token": "<|eot_id|>",
2056
+ "max_length": 1024,
2057
  "model_input_names": [
2058
  "input_ids",
2059
  "attention_mask"
2060
  ],
2061
  "model_max_length": 1000000000000000019884624838656,
2062
  "pad_token": "<|eot_id|>",
2063
+ "stride": 0,
2064
+ "tokenizer_class": "PreTrainedTokenizerFast",
2065
+ "truncation_side": "right",
2066
+ "truncation_strategy": "longest_first"
2067
  }
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:58d5bfb553da065051a173638d5415c4e5a09f001d12c1f1116fce4d4ccca8b4
3
  size 4667
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:596a0ab20874a6edffb19b3bcbfa30ef508de7c59c47f46a5205c74f5d3a5703
3
  size 4667