Hugging Face's logo

peterhung
/
EvolCodeLlama-7b

PEFT
Safetensors
llama
axolotl
Generated from Trainer
Model card Files Community
peterhung commited on
Commit
a7c6491
·
verified ·
1 Parent(s): 42a0dd8

End of training

Browse files
Files changed (2) hide show
  1. README.md +241 -0
  2. adapter_model.bin +3 -0
README.md ADDED
@@ -0,0 +1,241 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: llama2
3
+ library_name: peft
4
+ tags:
5
+ - axolotl
6
+ - generated_from_trainer
7
+ base_model: codellama/CodeLlama-7b-hf
8
+ model-index:
9
+ - name: EvolCodeLlama-7b
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
+ [<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)
17
+ <details><summary>See axolotl config</summary>
18
+
19
+ axolotl version: `0.4.0`
20
+ ```yaml
21
+ base_model: codellama/CodeLlama-7b-hf
22
+ base_model_config: codellama/CodeLlama-7b-hf
23
+ model_type: LlamaForCausalLM
24
+ tokenizer_type: LlamaTokenizer
25
+ is_llama_derived_model: true
26
+ hub_model_id: EvolCodeLlama-7b
27
+
28
+ load_in_8bit: false
29
+ load_in_4bit: true
30
+ strict: false
31
+
32
+ datasets:
33
+ - path: mlabonne/Evol-Instruct-Python-1k
34
+ type: alpaca
35
+ dataset_prepared_path: last_run_prepared
36
+ val_set_size: 0.02
37
+ output_dir: ./qlora-out
38
+
39
+ adapter: qlora
40
+ lora_model_dir:
41
+
42
+ sequence_len: 2048
43
+ sample_packing: true
44
+
45
+ lora_r: 32
46
+ lora_alpha: 16
47
+ lora_dropout: 0.05
48
+ lora_target_modules:
49
+ lora_target_linear: true
50
+ lora_fan_in_fan_out:
51
+
52
+ wandb_project: FTCodeLlama-2
53
+ wandb_entity:
54
+ wandb_watch:
55
+ wandb_run_id:
56
+ wandb_log_model:
57
+
58
+ gradient_accumulation_steps: 2
59
+ micro_batch_size: 4
60
+ num_epochs: 3
61
+ optimizer: paged_adamw_32bit
62
+ lr_scheduler: cosine
63
+ learning_rate: 0.0002
64
+
65
+ train_on_inputs: false
66
+ group_by_length: false
67
+ bf16: true
68
+ fp16: false
69
+ tf32: false
70
+
71
+ gradient_checkpointing: true
72
+ early_stopping_patience:
73
+ resume_from_checkpoint:
74
+ local_rank:
75
+ logging_steps: 1
76
+ xformers_attention:
77
+ flash_attention: true
78
+
79
+ warmup_steps: 100
80
+ eval_steps: 0.01
81
+ save_strategy: epoch
82
+ save_steps:
83
+ debug:
84
+ deepspeed:
85
+ weight_decay: 0.0
86
+ fsdp:
87
+ fsdp_config:
88
+ special_tokens:
89
+ bos_token: "<s>"
90
+ eos_token: "</s>"
91
+ unk_token: "<unk>"
92
+
93
+ ```
94
+
95
+ </details><br>
96
+
97
+ # EvolCodeLlama-7b
98
+
99
+ This model is a fine-tuned version of [codellama/CodeLlama-7b-hf](https://huggingface.co/codellama/CodeLlama-7b-hf) on the None dataset.
100
+ It achieves the following results on the evaluation set:
101
+ - Loss: 0.3754
102
+
103
+ ## Model description
104
+
105
+ More information needed
106
+
107
+ ## Intended uses & limitations
108
+
109
+ More information needed
110
+
111
+ ## Training and evaluation data
112
+
113
+ More information needed
114
+
115
+ ## Training procedure
116
+
117
+ ### Training hyperparameters
118
+
119
+ The following hyperparameters were used during training:
120
+ - learning_rate: 0.0002
121
+ - train_batch_size: 4
122
+ - eval_batch_size: 4
123
+ - seed: 42
124
+ - gradient_accumulation_steps: 2
125
+ - total_train_batch_size: 8
126
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
127
+ - lr_scheduler_type: cosine
128
+ - lr_scheduler_warmup_steps: 100
129
+ - num_epochs: 3
130
+
131
+ ### Training results
132
+
133
+ | Training Loss | Epoch | Step | Validation Loss |
134
+ |:-------------:|:-----:|:----:|:---------------:|
135
+ | 0.3686 | 0.01 | 1 | 0.5015 |
136
+ | 0.4397 | 0.03 | 3 | 0.5013 |
137
+ | 0.4919 | 0.06 | 6 | 0.5013 |
138
+ | 0.3191 | 0.09 | 9 | 0.5011 |
139
+ | 0.2514 | 0.12 | 12 | 0.5003 |
140
+ | 0.3379 | 0.15 | 15 | 0.4992 |
141
+ | 0.4712 | 0.19 | 18 | 0.4969 |
142
+ | 0.3801 | 0.22 | 21 | 0.4922 |
143
+ | 0.3482 | 0.25 | 24 | 0.4833 |
144
+ | 0.4113 | 0.28 | 27 | 0.4702 |
145
+ | 0.2524 | 0.31 | 30 | 0.4552 |
146
+ | 0.2641 | 0.34 | 33 | 0.4415 |
147
+ | 0.3554 | 0.37 | 36 | 0.4302 |
148
+ | 0.2384 | 0.4 | 39 | 0.4213 |
149
+ | 0.2131 | 0.43 | 42 | 0.4153 |
150
+ | 0.2308 | 0.46 | 45 | 0.4105 |
151
+ | 0.3478 | 0.49 | 48 | 0.4053 |
152
+ | 0.2913 | 0.53 | 51 | 0.4003 |
153
+ | 0.2909 | 0.56 | 54 | 0.3956 |
154
+ | 0.2032 | 0.59 | 57 | 0.3928 |
155
+ | 0.2479 | 0.62 | 60 | 0.3906 |
156
+ | 0.2145 | 0.65 | 63 | 0.3890 |
157
+ | 0.2447 | 0.68 | 66 | 0.3882 |
158
+ | 0.2928 | 0.71 | 69 | 0.3876 |
159
+ | 0.384 | 0.74 | 72 | 0.3854 |
160
+ | 0.1751 | 0.77 | 75 | 0.3835 |
161
+ | 0.352 | 0.8 | 78 | 0.3818 |
162
+ | 0.2443 | 0.84 | 81 | 0.3808 |
163
+ | 0.3211 | 0.87 | 84 | 0.3798 |
164
+ | 0.3026 | 0.9 | 87 | 0.3788 |
165
+ | 0.2357 | 0.93 | 90 | 0.3776 |
166
+ | 0.2661 | 0.96 | 93 | 0.3755 |
167
+ | 0.3314 | 0.99 | 96 | 0.3751 |
168
+ | 0.2789 | 1.02 | 99 | 0.3742 |
169
+ | 0.1734 | 1.03 | 102 | 0.3744 |
170
+ | 0.1928 | 1.06 | 105 | 0.3761 |
171
+ | 0.2681 | 1.09 | 108 | 0.3753 |
172
+ | 0.4148 | 1.12 | 111 | 0.3750 |
173
+ | 0.1977 | 1.15 | 114 | 0.3744 |
174
+ | 0.1977 | 1.19 | 117 | 0.3740 |
175
+ | 0.2499 | 1.22 | 120 | 0.3742 |
176
+ | 0.2192 | 1.25 | 123 | 0.3730 |
177
+ | 0.2207 | 1.28 | 126 | 0.3723 |
178
+ | 0.2179 | 1.31 | 129 | 0.3718 |
179
+ | 0.2843 | 1.34 | 132 | 0.3734 |
180
+ | 0.2614 | 1.37 | 135 | 0.3721 |
181
+ | 0.2033 | 1.4 | 138 | 0.3705 |
182
+ | 0.212 | 1.43 | 141 | 0.3705 |
183
+ | 0.2307 | 1.46 | 144 | 0.3712 |
184
+ | 0.3182 | 1.49 | 147 | 0.3698 |
185
+ | 0.2467 | 1.53 | 150 | 0.3664 |
186
+ | 0.1909 | 1.56 | 153 | 0.3665 |
187
+ | 0.3286 | 1.59 | 156 | 0.3655 |
188
+ | 0.2195 | 1.62 | 159 | 0.3648 |
189
+ | 0.3231 | 1.65 | 162 | 0.3650 |
190
+ | 0.2922 | 1.68 | 165 | 0.3631 |
191
+ | 0.1956 | 1.71 | 168 | 0.3627 |
192
+ | 0.2299 | 1.74 | 171 | 0.3639 |
193
+ | 0.1585 | 1.77 | 174 | 0.3649 |
194
+ | 0.2289 | 1.8 | 177 | 0.3650 |
195
+ | 0.189 | 1.84 | 180 | 0.3643 |
196
+ | 0.2736 | 1.87 | 183 | 0.3628 |
197
+ | 0.3591 | 1.9 | 186 | 0.3614 |
198
+ | 0.3181 | 1.93 | 189 | 0.3612 |
199
+ | 0.1994 | 1.96 | 192 | 0.3612 |
200
+ | 0.2499 | 1.99 | 195 | 0.3618 |
201
+ | 0.1659 | 2.01 | 198 | 0.3627 |
202
+ | 0.231 | 2.04 | 201 | 0.3665 |
203
+ | 0.169 | 2.07 | 204 | 0.3744 |
204
+ | 0.2082 | 2.1 | 207 | 0.3800 |
205
+ | 0.1755 | 2.13 | 210 | 0.3770 |
206
+ | 0.1959 | 2.16 | 213 | 0.3721 |
207
+ | 0.1933 | 2.19 | 216 | 0.3705 |
208
+ | 0.1213 | 2.22 | 219 | 0.3712 |
209
+ | 0.237 | 2.25 | 222 | 0.3738 |
210
+ | 0.2277 | 2.28 | 225 | 0.3771 |
211
+ | 0.2832 | 2.31 | 228 | 0.3789 |
212
+ | 0.2039 | 2.35 | 231 | 0.3783 |
213
+ | 0.2302 | 2.38 | 234 | 0.3764 |
214
+ | 0.1562 | 2.41 | 237 | 0.3750 |
215
+ | 0.1688 | 2.44 | 240 | 0.3742 |
216
+ | 0.126 | 2.47 | 243 | 0.3741 |
217
+ | 0.1846 | 2.5 | 246 | 0.3746 |
218
+ | 0.2195 | 2.53 | 249 | 0.3745 |
219
+ | 0.2335 | 2.56 | 252 | 0.3749 |
220
+ | 0.1542 | 2.59 | 255 | 0.3750 |
221
+ | 0.1783 | 2.62 | 258 | 0.3755 |
222
+ | 0.2409 | 2.65 | 261 | 0.3762 |
223
+ | 0.1777 | 2.69 | 264 | 0.3762 |
224
+ | 0.2591 | 2.72 | 267 | 0.3761 |
225
+ | 0.2092 | 2.75 | 270 | 0.3757 |
226
+ | 0.2256 | 2.78 | 273 | 0.3757 |
227
+ | 0.1923 | 2.81 | 276 | 0.3756 |
228
+ | 0.156 | 2.84 | 279 | 0.3755 |
229
+ | 0.2036 | 2.87 | 282 | 0.3754 |
230
+ | 0.2254 | 2.9 | 285 | 0.3753 |
231
+ | 0.1683 | 2.93 | 288 | 0.3753 |
232
+ | 0.1528 | 2.96 | 291 | 0.3754 |
233
+
234
+
235
+ ### Framework versions
236
+
237
+ - PEFT 0.8.2
238
+ - Transformers 4.38.0.dev0
239
+ - Pytorch 2.1.2+cu118
240
+ - Datasets 2.17.0
241
+ - Tokenizers 0.15.0
adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:00441eb9e92524fc94a56e496ce8ca0aa70c391158881a04807845486b7e0385
3
+ size 319977674