Momorami commited on
Commit
85266dc
1 Parent(s): c40b670

24/08/18 Initial Commit

Browse files
README.md CHANGED
@@ -1,3 +1,142 @@
1
- ---
2
- license: apache-2.0
3
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model: keeeeenw/MicroLlama
4
+ tags:
5
+ - generated_from_trainer
6
+ model-index:
7
+ - name: medusa-microllama_305M_stage2
8
+ results: []
9
+ ---
10
+
11
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
12
+ should probably proofread and complete it, then remove this comment. -->
13
+
14
+ [<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl)
15
+ [<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/momorami-kaist/medusa_test/runs/eg00n44l)
16
+ # medusa-microllama_305M_stage2
17
+
18
+ This model is a fine-tuned version of [keeeeenw/MicroLlama](https://huggingface.co/keeeeenw/MicroLlama) on the None dataset.
19
+ It achieves the following results on the evaluation set:
20
+ - Loss: 3.5262
21
+
22
+ ## Model description
23
+
24
+ More information needed
25
+
26
+ ## Intended uses & limitations
27
+
28
+ More information needed
29
+
30
+ ## Training and evaluation data
31
+
32
+ More information needed
33
+
34
+ ## Training procedure
35
+
36
+ ### Training hyperparameters
37
+
38
+ The following hyperparameters were used during training:
39
+ - learning_rate: 0.0005
40
+ - train_batch_size: 2
41
+ - eval_batch_size: 2
42
+ - seed: 42
43
+ - gradient_accumulation_steps: 4
44
+ - total_train_batch_size: 8
45
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
46
+ - lr_scheduler_type: cosine
47
+ - lr_scheduler_warmup_steps: 40
48
+ - num_epochs: 2
49
+
50
+ ### Training results
51
+
52
+ | Training Loss | Epoch | Step | Validation Loss |
53
+ |:-------------:|:------:|:----:|:---------------:|
54
+ | 4.6913 | 0.0244 | 40 | 4.7578 |
55
+ | 4.8782 | 0.0489 | 80 | 4.8017 |
56
+ | 4.642 | 0.0733 | 120 | 4.7973 |
57
+ | 4.4601 | 0.0978 | 160 | 4.7589 |
58
+ | 4.4806 | 0.1222 | 200 | 4.6955 |
59
+ | 4.4856 | 0.1467 | 240 | 4.6196 |
60
+ | 4.4671 | 0.1711 | 280 | 4.5750 |
61
+ | 4.3228 | 0.1955 | 320 | 4.5563 |
62
+ | 4.1184 | 0.2200 | 360 | 4.5274 |
63
+ | 3.9986 | 0.2444 | 400 | 4.5031 |
64
+ | 4.2603 | 0.2689 | 440 | 4.4637 |
65
+ | 4.1344 | 0.2933 | 480 | 4.4349 |
66
+ | 4.1973 | 0.3178 | 520 | 4.4106 |
67
+ | 4.3961 | 0.3422 | 560 | 4.4202 |
68
+ | 4.1814 | 0.3666 | 600 | 4.3732 |
69
+ | 4.1685 | 0.3911 | 640 | 4.3877 |
70
+ | 4.3108 | 0.4155 | 680 | 4.3262 |
71
+ | 4.6294 | 0.4400 | 720 | 4.3108 |
72
+ | 4.3653 | 0.4644 | 760 | 4.2880 |
73
+ | 4.1505 | 0.4888 | 800 | 4.2835 |
74
+ | 3.8278 | 0.5133 | 840 | 4.2623 |
75
+ | 4.3567 | 0.5377 | 880 | 4.2253 |
76
+ | 4.2782 | 0.5622 | 920 | 4.1919 |
77
+ | 4.1025 | 0.5866 | 960 | 4.1846 |
78
+ | 4.2819 | 0.6111 | 1000 | 4.1637 |
79
+ | 3.9919 | 0.6355 | 1040 | 4.1323 |
80
+ | 4.1932 | 0.6599 | 1080 | 4.1017 |
81
+ | 4.0949 | 0.6844 | 1120 | 4.1085 |
82
+ | 3.7266 | 0.7088 | 1160 | 4.0668 |
83
+ | 4.1255 | 0.7333 | 1200 | 4.0500 |
84
+ | 4.3707 | 0.7577 | 1240 | 4.0207 |
85
+ | 4.1965 | 0.7822 | 1280 | 4.0065 |
86
+ | 3.4585 | 0.8066 | 1320 | 3.9363 |
87
+ | 3.7242 | 0.8310 | 1360 | 3.8893 |
88
+ | 3.9228 | 0.8555 | 1400 | 3.8569 |
89
+ | 4.2051 | 0.8799 | 1440 | 3.8412 |
90
+ | 3.6795 | 0.9044 | 1480 | 3.8245 |
91
+ | 3.2453 | 0.9288 | 1520 | 3.8132 |
92
+ | 3.5941 | 0.9533 | 1560 | 3.7907 |
93
+ | 3.6246 | 0.9777 | 1600 | 3.7573 |
94
+ | 2.8637 | 1.0021 | 1640 | 3.7530 |
95
+ | 2.8495 | 1.0266 | 1680 | 3.7741 |
96
+ | 3.0246 | 1.0510 | 1720 | 3.7690 |
97
+ | 2.99 | 1.0755 | 1760 | 3.7464 |
98
+ | 3.1902 | 1.0999 | 1800 | 3.7347 |
99
+ | 2.8099 | 1.1244 | 1840 | 3.7278 |
100
+ | 2.7652 | 1.1488 | 1880 | 3.7245 |
101
+ | 2.6362 | 1.1732 | 1920 | 3.7034 |
102
+ | 2.8562 | 1.1977 | 1960 | 3.6871 |
103
+ | 3.1712 | 1.2221 | 2000 | 3.6786 |
104
+ | 2.7405 | 1.2466 | 2040 | 3.6709 |
105
+ | 2.734 | 1.2710 | 2080 | 3.6404 |
106
+ | 3.1788 | 1.2954 | 2120 | 3.6310 |
107
+ | 2.9609 | 1.3199 | 2160 | 3.6176 |
108
+ | 3.0737 | 1.3443 | 2200 | 3.6136 |
109
+ | 2.751 | 1.3688 | 2240 | 3.5960 |
110
+ | 2.7105 | 1.3932 | 2280 | 3.5872 |
111
+ | 2.8158 | 1.4177 | 2320 | 3.5848 |
112
+ | 3.03 | 1.4421 | 2360 | 3.5679 |
113
+ | 2.8122 | 1.4665 | 2400 | 3.5718 |
114
+ | 2.5581 | 1.4910 | 2440 | 3.5568 |
115
+ | 2.9845 | 1.5154 | 2480 | 3.5496 |
116
+ | 2.83 | 1.5399 | 2520 | 3.5440 |
117
+ | 2.7004 | 1.5643 | 2560 | 3.5402 |
118
+ | 2.8271 | 1.5888 | 2600 | 3.5406 |
119
+ | 2.5315 | 1.6132 | 2640 | 3.5316 |
120
+ | 2.6001 | 1.6376 | 2680 | 3.5346 |
121
+ | 2.4959 | 1.6621 | 2720 | 3.5298 |
122
+ | 2.9174 | 1.6865 | 2760 | 3.5304 |
123
+ | 2.7219 | 1.7110 | 2800 | 3.5286 |
124
+ | 2.5395 | 1.7354 | 2840 | 3.5279 |
125
+ | 2.7464 | 1.7599 | 2880 | 3.5284 |
126
+ | 2.7532 | 1.7843 | 2920 | 3.5274 |
127
+ | 2.6472 | 1.8087 | 2960 | 3.5270 |
128
+ | 2.8263 | 1.8332 | 3000 | 3.5268 |
129
+ | 2.916 | 1.8576 | 3040 | 3.5263 |
130
+ | 3.0202 | 1.8821 | 3080 | 3.5262 |
131
+ | 2.7152 | 1.9065 | 3120 | 3.5261 |
132
+ | 2.7628 | 1.9310 | 3160 | 3.5261 |
133
+ | 2.783 | 1.9554 | 3200 | 3.5263 |
134
+ | 3.2587 | 1.9798 | 3240 | 3.5262 |
135
+
136
+
137
+ ### Framework versions
138
+
139
+ - Transformers 4.43.0
140
+ - Pytorch 2.3.1
141
+ - Datasets 2.15.0
142
+ - Tokenizers 0.19.1
config.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "keeeeenw/MicroLlama",
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": 1024,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 5632,
14
+ "max_position_embeddings": 2048,
15
+ "medusa_num_heads": 4,
16
+ "medusa_num_layers": 1,
17
+ "mlp_bias": false,
18
+ "model_type": "llama",
19
+ "num_attention_heads": 16,
20
+ "num_hidden_layers": 12,
21
+ "num_key_value_heads": 4,
22
+ "pretraining_tp": 1,
23
+ "rms_norm_eps": 1e-05,
24
+ "rope_scaling": null,
25
+ "rope_theta": 10000.0,
26
+ "tie_word_embeddings": false,
27
+ "torch_dtype": "bfloat16",
28
+ "transformers_version": "4.43.0",
29
+ "use_cache": false,
30
+ "vocab_size": 32000
31
+ }
generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "transformers_version": "4.43.0"
6
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:be4eb34ca0e3fc34f34baac894f0f44680a3f058668a92ad29c0e762bf753f32
3
+ size 879828432
optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:30135acbe7354f09dd92c92216071e9d4d877727c1b50266f08977b59f3a744a
3
+ size 881840890
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:742702a217271b44f6ba575b260d03262a496587d84b8b63e8e3f4397032ac7c
3
+ size 879855990
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9196a1e708bf24d6abba41cce3f8558820acc3e50f9394c5955e29eb41ffea3d
3
+ size 14244
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8039d280dad3070e169b806490ff8e0bd5e1077911a754b77ef2909860fde236
3
+ size 1064
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": "</s>",
17
+ "unk_token": {
18
+ "content": "<unk>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
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 +1,43 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": true,
5
+ "added_tokens_decoder": {
6
+ "0": {
7
+ "content": "<unk>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false,
12
+ "special": true
13
+ },
14
+ "1": {
15
+ "content": "<s>",
16
+ "lstrip": false,
17
+ "normalized": false,
18
+ "rstrip": false,
19
+ "single_word": false,
20
+ "special": true
21
+ },
22
+ "2": {
23
+ "content": "</s>",
24
+ "lstrip": false,
25
+ "normalized": false,
26
+ "rstrip": false,
27
+ "single_word": false,
28
+ "special": true
29
+ }
30
+ },
31
+ "bos_token": "<s>",
32
+ "clean_up_tokenization_spaces": false,
33
+ "eos_token": "</s>",
34
+ "legacy": true,
35
+ "model_max_length": 2048,
36
+ "pad_token": "</s>",
37
+ "sp_model_kwargs": {},
38
+ "spaces_between_special_tokens": false,
39
+ "tokenizer_class": "LlamaTokenizer",
40
+ "unk_token": "<unk>",
41
+ "use_default_system_prompt": false,
42
+ "use_fast": true
43
+ }
trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9fe8eb74cf7eb4ffabe9c2817f4ff602404e640fa2f457eb78e8f5a5d58eab00
3
+ size 5624