martimfasantos commited on
Commit
ec8a6e7
1 Parent(s): d635470

End of training

Browse files
README.md ADDED
@@ -0,0 +1,98 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model: TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T
4
+ tags:
5
+ - xcomet_xl_xxl
6
+ - generated_from_trainer
7
+ model-index:
8
+ - name: cpo-xcomet-xl_xxl-inc7b-10p-shuff-1e-7-full-tiny
9
+ results: []
10
+ ---
11
+
12
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
13
+ should probably proofread and complete it, then remove this comment. -->
14
+
15
+ # cpo-xcomet-xl_xxl-inc7b-10p-shuff-1e-7-full-tiny
16
+
17
+ This model is a fine-tuned version of [TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T](https://huggingface.co/TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T) on the Unbabel/TowerAligned-v0.1 dataset.
18
+ It achieves the following results on the evaluation set:
19
+ - Loss: 2.7960
20
+ - Nll Loss: 1.0602
21
+ - Logps/best: -102.8461
22
+ - Rewards/chosen: -10.2846
23
+ - Rewards/rejected: -9.6988
24
+ - Rewards/accuracies: 0.4600
25
+ - Rewards/margins: -0.5858
26
+ - Logps/rejected: -96.9882
27
+ - Logps/chosen: -102.8461
28
+ - Logits/rejected: -1.8264
29
+ - Logits/chosen: -1.9635
30
+
31
+ ## Model description
32
+
33
+ More information needed
34
+
35
+ ## Intended uses & limitations
36
+
37
+ More information needed
38
+
39
+ ## Training and evaluation data
40
+
41
+ More information needed
42
+
43
+ ## Training procedure
44
+
45
+ ### Training hyperparameters
46
+
47
+ The following hyperparameters were used during training:
48
+ - learning_rate: 1e-07
49
+ - train_batch_size: 1
50
+ - eval_batch_size: 4
51
+ - seed: 42
52
+ - gradient_accumulation_steps: 16
53
+ - total_train_batch_size: 16
54
+ - optimizer: Adam with betas=(0.9,0.95) and epsilon=1e-08
55
+ - lr_scheduler_type: linear
56
+ - lr_scheduler_warmup_ratio: 0.1
57
+ - num_epochs: 3
58
+
59
+ ### Training results
60
+
61
+ | Training Loss | Epoch | Step | Validation Loss | Nll Loss | Logps/best | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
62
+ |:-------------:|:------:|:----:|:---------------:|:--------:|:----------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
63
+ | 2.9687 | 0.1063 | 100 | 2.8066 | 1.0659 | -103.3585 | -10.3358 | -9.7431 | 0.4540 | -0.5928 | -97.4308 | -103.3585 | -1.8274 | -1.9646 |
64
+ | 3.0173 | 0.2127 | 200 | 2.8063 | 1.0656 | -103.3396 | -10.3340 | -9.7402 | 0.4560 | -0.5937 | -97.4022 | -103.3396 | -1.8275 | -1.9648 |
65
+ | 2.8267 | 0.3190 | 300 | 2.8046 | 1.0650 | -103.2849 | -10.3285 | -9.7373 | 0.4540 | -0.5912 | -97.3725 | -103.2849 | -1.8273 | -1.9644 |
66
+ | 2.9404 | 0.4254 | 400 | 2.8046 | 1.0644 | -103.2318 | -10.3232 | -9.7301 | 0.4600 | -0.5931 | -97.3013 | -103.2318 | -1.8271 | -1.9643 |
67
+ | 3.3065 | 0.5317 | 500 | 2.8002 | 1.0637 | -103.1556 | -10.3156 | -9.7280 | 0.4600 | -0.5875 | -97.2803 | -103.1556 | -1.8268 | -1.9640 |
68
+ | 2.9333 | 0.6381 | 600 | 2.8021 | 1.0633 | -103.1282 | -10.3128 | -9.7212 | 0.4560 | -0.5916 | -97.2122 | -103.1282 | -1.8271 | -1.9642 |
69
+ | 3.2698 | 0.7444 | 700 | 2.8006 | 1.0627 | -103.0742 | -10.3074 | -9.7178 | 0.4580 | -0.5897 | -97.1777 | -103.0742 | -1.8268 | -1.9640 |
70
+ | 2.7002 | 0.8508 | 800 | 2.8003 | 1.0624 | -103.0458 | -10.3046 | -9.7147 | 0.4580 | -0.5899 | -97.1470 | -103.0458 | -1.8269 | -1.9641 |
71
+ | 3.0848 | 0.9571 | 900 | 2.7984 | 1.0620 | -103.0023 | -10.3002 | -9.7132 | 0.4580 | -0.5870 | -97.1324 | -103.0023 | -1.8267 | -1.9638 |
72
+ | 2.9243 | 1.0635 | 1000 | 2.7987 | 1.0617 | -102.9805 | -10.2980 | -9.7086 | 0.4580 | -0.5895 | -97.0859 | -102.9805 | -1.8268 | -1.9639 |
73
+ | 2.7945 | 1.1698 | 1100 | 2.7974 | 1.0615 | -102.9564 | -10.2956 | -9.7084 | 0.4580 | -0.5872 | -97.0842 | -102.9564 | -1.8267 | -1.9638 |
74
+ | 2.7893 | 1.2762 | 1200 | 2.7979 | 1.0613 | -102.9413 | -10.2941 | -9.7061 | 0.4620 | -0.5880 | -97.0609 | -102.9413 | -1.8266 | -1.9638 |
75
+ | 3.2162 | 1.3825 | 1300 | 2.7978 | 1.0611 | -102.9208 | -10.2921 | -9.7039 | 0.4540 | -0.5882 | -97.0387 | -102.9208 | -1.8266 | -1.9637 |
76
+ | 2.8123 | 1.4889 | 1400 | 2.7980 | 1.0611 | -102.9247 | -10.2925 | -9.7032 | 0.4580 | -0.5893 | -97.0320 | -102.9247 | -1.8266 | -1.9637 |
77
+ | 2.785 | 1.5952 | 1500 | 2.7973 | 1.0606 | -102.8798 | -10.2880 | -9.6993 | 0.4560 | -0.5887 | -96.9928 | -102.8798 | -1.8265 | -1.9636 |
78
+ | 2.7997 | 1.7016 | 1600 | 2.7952 | 1.0606 | -102.8751 | -10.2875 | -9.7026 | 0.4600 | -0.5849 | -97.0257 | -102.8751 | -1.8267 | -1.9638 |
79
+ | 2.6655 | 1.8079 | 1700 | 2.7956 | 1.0605 | -102.8628 | -10.2863 | -9.7005 | 0.4620 | -0.5858 | -97.0050 | -102.8628 | -1.8264 | -1.9635 |
80
+ | 2.7597 | 1.9143 | 1800 | 2.7966 | 1.0605 | -102.8715 | -10.2871 | -9.6999 | 0.4540 | -0.5872 | -96.9991 | -102.8715 | -1.8267 | -1.9637 |
81
+ | 2.9736 | 2.0206 | 1900 | 2.7955 | 1.0603 | -102.8511 | -10.2851 | -9.6990 | 0.4600 | -0.5861 | -96.9900 | -102.8511 | -1.8266 | -1.9637 |
82
+ | 2.8977 | 2.1270 | 2000 | 2.7954 | 1.0603 | -102.8550 | -10.2855 | -9.6990 | 0.4560 | -0.5865 | -96.9901 | -102.8550 | -1.8270 | -1.9641 |
83
+ | 2.7043 | 2.2333 | 2100 | 2.7961 | 1.0604 | -102.8632 | -10.2863 | -9.6997 | 0.4560 | -0.5867 | -96.9967 | -102.8632 | -1.8264 | -1.9635 |
84
+ | 2.7693 | 2.3396 | 2200 | 2.7951 | 1.0604 | -102.8550 | -10.2855 | -9.6998 | 0.4600 | -0.5857 | -96.9983 | -102.8550 | -1.8263 | -1.9634 |
85
+ | 2.6632 | 2.4460 | 2300 | 2.7943 | 1.0602 | -102.8407 | -10.2841 | -9.6989 | 0.4600 | -0.5851 | -96.9893 | -102.8407 | -1.8264 | -1.9635 |
86
+ | 3.2451 | 2.5523 | 2400 | 2.7953 | 1.0602 | -102.8434 | -10.2843 | -9.6989 | 0.4580 | -0.5855 | -96.9885 | -102.8434 | -1.8264 | -1.9635 |
87
+ | 2.7117 | 2.6587 | 2500 | 2.7955 | 1.0601 | -102.8357 | -10.2836 | -9.6962 | 0.4580 | -0.5873 | -96.9625 | -102.8357 | -1.8263 | -1.9634 |
88
+ | 3.148 | 2.7650 | 2600 | 2.7967 | 1.0604 | -102.8636 | -10.2864 | -9.6985 | 0.4560 | -0.5878 | -96.9853 | -102.8636 | -1.8265 | -1.9636 |
89
+ | 3.2951 | 2.8714 | 2700 | 2.7959 | 1.0602 | -102.8490 | -10.2849 | -9.6981 | 0.4620 | -0.5868 | -96.9812 | -102.8490 | -1.8263 | -1.9634 |
90
+ | 2.8486 | 2.9777 | 2800 | 2.7960 | 1.0602 | -102.8461 | -10.2846 | -9.6988 | 0.4600 | -0.5858 | -96.9882 | -102.8461 | -1.8264 | -1.9635 |
91
+
92
+
93
+ ### Framework versions
94
+
95
+ - Transformers 4.41.2
96
+ - Pytorch 2.1.2
97
+ - Datasets 2.20.0
98
+ - Tokenizers 0.19.1
config.json ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T",
3
+ "architectures": [
4
+ "LlamaForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "bos_token_id": 1,
9
+ "eos_token_id": 2,
10
+ "hidden_act": "silu",
11
+ "hidden_size": 2048,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 5632,
14
+ "max_position_embeddings": 2048,
15
+ "mlp_bias": false,
16
+ "model_type": "llama",
17
+ "num_attention_heads": 32,
18
+ "num_hidden_layers": 22,
19
+ "num_key_value_heads": 4,
20
+ "pretraining_tp": 1,
21
+ "rms_norm_eps": 1e-05,
22
+ "rope_scaling": null,
23
+ "rope_theta": 10000.0,
24
+ "tie_word_embeddings": false,
25
+ "torch_dtype": "bfloat16",
26
+ "transformers_version": "4.41.2",
27
+ "use_cache": true,
28
+ "vocab_size": 32000
29
+ }
generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 1,
3
+ "eos_token_id": 2,
4
+ "max_length": 2048,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.41.2"
7
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aaba0e8afe6ec337e6de464697267d74a6d5447faafce8d6f98e4835b3bdd8dd
3
+ size 2200119864
special_tokens_map.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "[PAD]",
17
+ "unk_token": {
18
+ "content": "<unk>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
tokenizer_config.json ADDED
@@ -0,0 +1,41 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "added_tokens_decoder": {
5
+ "0": {
6
+ "content": "<unk>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "1": {
14
+ "content": "<s>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "2": {
22
+ "content": "</s>",
23
+ "lstrip": false,
24
+ "normalized": false,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ }
29
+ },
30
+ "bos_token": "<s>",
31
+ "clean_up_tokenization_spaces": false,
32
+ "eos_token": "</s>",
33
+ "legacy": false,
34
+ "model_max_length": 1000000000000000019884624838656,
35
+ "pad_token": "[PAD]",
36
+ "padding_side": "left",
37
+ "sp_model_kwargs": {},
38
+ "tokenizer_class": "LlamaTokenizer",
39
+ "unk_token": "<unk>",
40
+ "use_default_system_prompt": false
41
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0dbd442b451f333522d2ec4c6768e395dec10eb58001851057f66a124a0f3c7d
3
+ size 5240