Hugging Face's logo

willtensora
/
a7f83208-aa47-4ecd-80e6-f21bda70bb90

PEFT
Safetensors
qwen2
axolotl
Generated from Trainer
Model card Files
xet
Community
1

Improve language tag

#1
by lbourdois - opened
base: refs/heads/main
←
from: refs/pr/1
Discussion Files changed
+242
-228
Files changed (1) hide show
  1. README.md +242 -228
README.md CHANGED
@@ -1,229 +1,243 @@
1
- ---
2
- library_name: peft
3
- license: apache-2.0
4
- base_model: Qwen/Qwen2.5-0.5B
5
- tags:
6
- - axolotl
7
- - generated_from_trainer
8
- model-index:
9
- - name: a7f83208-aa47-4ecd-80e6-f21bda70bb90
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/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl)
17
- <details><summary>See axolotl config</summary>
18
-
19
- axolotl version: `0.4.1`
20
- ```yaml
21
- adapter: lora
22
- base_model: Qwen/Qwen2.5-0.5B
23
- batch_size: 8
24
- bf16: true
25
- chat_template: tokenizer_default_fallback_alpaca
26
- datasets:
27
- - data_files:
28
- - 44664facd5408a4c_train_data.json
29
- ds_type: json
30
- format: custom
31
- path: /workspace/input_data/44664facd5408a4c_train_data.json
32
- type:
33
- field_input: choices
34
- field_instruction: full_prompt
35
- field_output: example
36
- format: '{instruction} {input}'
37
- no_input_format: '{instruction}'
38
- system_format: '{system}'
39
- system_prompt: ''
40
- evals_per_epoch: 1
41
- flash_attention: true
42
- gpu_memory_limit: 80GiB
43
- gradient_checkpointing: true
44
- group_by_length: true
45
- hub_model_id: willtensora/a7f83208-aa47-4ecd-80e6-f21bda70bb90
46
- hub_strategy: checkpoint
47
- learning_rate: 0.0002
48
- logging_steps: 10
49
- lora_alpha: 256
50
- lora_dropout: 0.1
51
- lora_r: 128
52
- lora_target_linear: true
53
- lr_scheduler: cosine
54
- micro_batch_size: 1
55
- model_type: AutoModelForCausalLM
56
- num_epochs: 100
57
- optimizer: adamw_bnb_8bit
58
- output_dir: miner_id_24
59
- pad_to_sequence_len: true
60
- resize_token_embeddings_to_32x: false
61
- sample_packing: false
62
- saves_per_epoch: 2
63
- sequence_len: 2048
64
- tokenizer_type: Qwen2TokenizerFast
65
- train_on_inputs: false
66
- trust_remote_code: true
67
- val_set_size: 0.1
68
- wandb_entity: ''
69
- wandb_mode: online
70
- wandb_project: Gradients-On-Demand
71
- wandb_run: your_name
72
- wandb_runid: default
73
- warmup_ratio: 0.05
74
- xformers_attention: true
75
-
76
- ```
77
-
78
- </details><br>
79
-
80
- # a7f83208-aa47-4ecd-80e6-f21bda70bb90
81
-
82
- This model is a fine-tuned version of [Qwen/Qwen2.5-0.5B](https://huggingface.co/Qwen/Qwen2.5-0.5B) on the None dataset.
83
- It achieves the following results on the evaluation set:
84
- - Loss: 0.0000
85
-
86
- ## Model description
87
-
88
- More information needed
89
-
90
- ## Intended uses & limitations
91
-
92
- More information needed
93
-
94
- ## Training and evaluation data
95
-
96
- More information needed
97
-
98
- ## Training procedure
99
-
100
- ### Training hyperparameters
101
-
102
- The following hyperparameters were used during training:
103
- - learning_rate: 0.0002
104
- - train_batch_size: 1
105
- - eval_batch_size: 1
106
- - seed: 42
107
- - distributed_type: multi-GPU
108
- - num_devices: 8
109
- - total_train_batch_size: 8
110
- - total_eval_batch_size: 8
111
- - optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
112
- - lr_scheduler_type: cosine
113
- - lr_scheduler_warmup_steps: 24
114
- - num_epochs: 100
115
-
116
- ### Training results
117
-
118
- | Training Loss | Epoch | Step | Validation Loss |
119
- |:-------------:|:-----:|:----:|:---------------:|
120
- | No log | 0.025 | 1 | 0.9385 |
121
- | 0.0292 | 1.0 | 40 | 0.0043 |
122
- | 0.0148 | 2.0 | 80 | 0.0332 |
123
- | 0.1015 | 3.0 | 120 | 0.0044 |
124
- | 0.0002 | 4.0 | 160 | 0.0001 |
125
- | 0.0 | 5.0 | 200 | 0.0000 |
126
- | 0.0 | 6.0 | 240 | 0.0000 |
127
- | 0.0 | 7.0 | 280 | 0.0000 |
128
- | 0.0 | 8.0 | 320 | 0.0000 |
129
- | 0.0 | 9.0 | 360 | 0.0000 |
130
- | 0.0 | 10.0 | 400 | 0.0000 |
131
- | 0.0 | 11.0 | 440 | 0.0000 |
132
- | 0.0 | 12.0 | 480 | 0.0000 |
133
- | 0.0 | 13.0 | 520 | 0.0000 |
134
- | 0.0 | 14.0 | 560 | 0.0000 |
135
- | 0.0 | 15.0 | 600 | 0.0000 |
136
- | 0.0 | 16.0 | 640 | 0.0000 |
137
- | 0.0 | 17.0 | 680 | 0.0000 |
138
- | 0.0 | 18.0 | 720 | 0.0000 |
139
- | 0.0 | 19.0 | 760 | 0.0000 |
140
- | 0.0 | 20.0 | 800 | 0.0000 |
141
- | 0.0 | 21.0 | 840 | 0.0000 |
142
- | 0.0 | 22.0 | 880 | 0.0000 |
143
- | 0.0 | 23.0 | 920 | 0.0000 |
144
- | 0.0 | 24.0 | 960 | 0.0000 |
145
- | 0.0 | 25.0 | 1000 | 0.0000 |
146
- | 0.0 | 26.0 | 1040 | 0.0000 |
147
- | 0.0 | 27.0 | 1080 | 0.0000 |
148
- | 0.0 | 28.0 | 1120 | 0.0000 |
149
- | 0.0 | 29.0 | 1160 | 0.0000 |
150
- | 0.0 | 30.0 | 1200 | 0.0000 |
151
- | 0.0 | 31.0 | 1240 | 0.0000 |
152
- | 0.0 | 32.0 | 1280 | 0.0000 |
153
- | 0.0 | 33.0 | 1320 | 0.0000 |
154
- | 0.0 | 34.0 | 1360 | 0.0000 |
155
- | 0.0 | 35.0 | 1400 | 0.0000 |
156
- | 0.0 | 36.0 | 1440 | 0.0000 |
157
- | 0.0 | 37.0 | 1480 | 0.0000 |
158
- | 0.0 | 38.0 | 1520 | 0.0000 |
159
- | 0.0 | 39.0 | 1560 | 0.0000 |
160
- | 0.0 | 40.0 | 1600 | 0.0000 |
161
- | 0.0 | 41.0 | 1640 | 0.0000 |
162
- | 0.0 | 42.0 | 1680 | 0.0000 |
163
- | 0.0 | 43.0 | 1720 | 0.0000 |
164
- | 0.0 | 44.0 | 1760 | 0.0000 |
165
- | 0.0 | 45.0 | 1800 | 0.0000 |
166
- | 0.0 | 46.0 | 1840 | 0.0000 |
167
- | 0.0 | 47.0 | 1880 | 0.0000 |
168
- | 0.0 | 48.0 | 1920 | 0.0000 |
169
- | 0.0 | 49.0 | 1960 | 0.0000 |
170
- | 0.0 | 50.0 | 2000 | 0.0000 |
171
- | 0.0 | 51.0 | 2040 | 0.0000 |
172
- | 0.0 | 52.0 | 2080 | 0.0000 |
173
- | 0.0 | 53.0 | 2120 | 0.0000 |
174
- | 0.0 | 54.0 | 2160 | 0.0000 |
175
- | 0.0 | 55.0 | 2200 | 0.0000 |
176
- | 0.0 | 56.0 | 2240 | 0.0000 |
177
- | 0.0 | 57.0 | 2280 | 0.0000 |
178
- | 0.0 | 58.0 | 2320 | 0.0000 |
179
- | 0.0 | 59.0 | 2360 | 0.0000 |
180
- | 0.0 | 60.0 | 2400 | 0.0000 |
181
- | 0.0 | 61.0 | 2440 | 0.0000 |
182
- | 0.0 | 62.0 | 2480 | 0.0000 |
183
- | 0.0 | 63.0 | 2520 | 0.0000 |
184
- | 0.0 | 64.0 | 2560 | 0.0000 |
185
- | 0.0 | 65.0 | 2600 | 0.0000 |
186
- | 0.0 | 66.0 | 2640 | 0.0000 |
187
- | 0.0 | 67.0 | 2680 | 0.0000 |
188
- | 0.0 | 68.0 | 2720 | 0.0000 |
189
- | 0.0 | 69.0 | 2760 | 0.0000 |
190
- | 0.0 | 70.0 | 2800 | 0.0000 |
191
- | 0.0 | 71.0 | 2840 | 0.0000 |
192
- | 0.0 | 72.0 | 2880 | 0.0000 |
193
- | 0.0 | 73.0 | 2920 | 0.0000 |
194
- | 0.0 | 74.0 | 2960 | 0.0000 |
195
- | 0.0 | 75.0 | 3000 | 0.0000 |
196
- | 0.0 | 76.0 | 3040 | 0.0000 |
197
- | 0.0 | 77.0 | 3080 | 0.0000 |
198
- | 0.0 | 78.0 | 3120 | 0.0000 |
199
- | 0.0 | 79.0 | 3160 | 0.0000 |
200
- | 0.0 | 80.0 | 3200 | 0.0000 |
201
- | 0.0 | 81.0 | 3240 | 0.0000 |
202
- | 0.0 | 82.0 | 3280 | 0.0000 |
203
- | 0.0 | 83.0 | 3320 | 0.0000 |
204
- | 0.0 | 84.0 | 3360 | 0.0000 |
205
- | 0.0 | 85.0 | 3400 | 0.0000 |
206
- | 0.0 | 86.0 | 3440 | 0.0000 |
207
- | 0.0 | 87.0 | 3480 | 0.0000 |
208
- | 0.0 | 88.0 | 3520 | 0.0000 |
209
- | 0.0 | 89.0 | 3560 | 0.0000 |
210
- | 0.0 | 90.0 | 3600 | 0.0000 |
211
- | 0.0 | 91.0 | 3640 | 0.0000 |
212
- | 0.0 | 92.0 | 3680 | 0.0000 |
213
- | 0.0 | 93.0 | 3720 | 0.0000 |
214
- | 0.0 | 94.0 | 3760 | 0.0000 |
215
- | 0.0 | 95.0 | 3800 | 0.0000 |
216
- | 0.0 | 96.0 | 3840 | 0.0000 |
217
- | 0.0 | 97.0 | 3880 | 0.0000 |
218
- | 0.0 | 98.0 | 3920 | 0.0000 |
219
- | 0.0 | 99.0 | 3960 | 0.0000 |
220
- | 0.0 | 100.0 | 4000 | 0.0000 |
221
-
222
-
223
- ### Framework versions
224
-
225
- - PEFT 0.13.2
226
- - Transformers 4.46.0
227
- - Pytorch 2.5.0+cu124
228
- - Datasets 3.0.1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
229
  - Tokenizers 0.20.1
 
1
+ ---
2
+ library_name: peft
3
+ license: apache-2.0
4
+ base_model: Qwen/Qwen2.5-0.5B
5
+ tags:
6
+ - axolotl
7
+ - generated_from_trainer
8
+ language:
9
+ - zho
10
+ - eng
11
+ - fra
12
+ - spa
13
+ - por
14
+ - deu
15
+ - ita
16
+ - rus
17
+ - jpn
18
+ - kor
19
+ - vie
20
+ - tha
21
+ - ara
22
+ model-index:
23
+ - name: a7f83208-aa47-4ecd-80e6-f21bda70bb90
24
+ results: []
25
+ ---
26
+
27
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
28
+ should probably proofread and complete it, then remove this comment. -->
29
+
30
+ [<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl)
31
+ <details><summary>See axolotl config</summary>
32
+
33
+ axolotl version: `0.4.1`
34
+ ```yaml
35
+ adapter: lora
36
+ base_model: Qwen/Qwen2.5-0.5B
37
+ batch_size: 8
38
+ bf16: true
39
+ chat_template: tokenizer_default_fallback_alpaca
40
+ datasets:
41
+ - data_files:
42
+ - 44664facd5408a4c_train_data.json
43
+ ds_type: json
44
+ format: custom
45
+ path: /workspace/input_data/44664facd5408a4c_train_data.json
46
+ type:
47
+ field_input: choices
48
+ field_instruction: full_prompt
49
+ field_output: example
50
+ format: '{instruction} {input}'
51
+ no_input_format: '{instruction}'
52
+ system_format: '{system}'
53
+ system_prompt: ''
54
+ evals_per_epoch: 1
55
+ flash_attention: true
56
+ gpu_memory_limit: 80GiB
57
+ gradient_checkpointing: true
58
+ group_by_length: true
59
+ hub_model_id: willtensora/a7f83208-aa47-4ecd-80e6-f21bda70bb90
60
+ hub_strategy: checkpoint
61
+ learning_rate: 0.0002
62
+ logging_steps: 10
63
+ lora_alpha: 256
64
+ lora_dropout: 0.1
65
+ lora_r: 128
66
+ lora_target_linear: true
67
+ lr_scheduler: cosine
68
+ micro_batch_size: 1
69
+ model_type: AutoModelForCausalLM
70
+ num_epochs: 100
71
+ optimizer: adamw_bnb_8bit
72
+ output_dir: miner_id_24
73
+ pad_to_sequence_len: true
74
+ resize_token_embeddings_to_32x: false
75
+ sample_packing: false
76
+ saves_per_epoch: 2
77
+ sequence_len: 2048
78
+ tokenizer_type: Qwen2TokenizerFast
79
+ train_on_inputs: false
80
+ trust_remote_code: true
81
+ val_set_size: 0.1
82
+ wandb_entity: ''
83
+ wandb_mode: online
84
+ wandb_project: Gradients-On-Demand
85
+ wandb_run: your_name
86
+ wandb_runid: default
87
+ warmup_ratio: 0.05
88
+ xformers_attention: true
89
+
90
+ ```
91
+
92
+ </details><br>
93
+
94
+ # a7f83208-aa47-4ecd-80e6-f21bda70bb90
95
+
96
+ This model is a fine-tuned version of [Qwen/Qwen2.5-0.5B](https://huggingface.co/Qwen/Qwen2.5-0.5B) on the None dataset.
97
+ It achieves the following results on the evaluation set:
98
+ - Loss: 0.0000
99
+
100
+ ## Model description
101
+
102
+ More information needed
103
+
104
+ ## Intended uses & limitations
105
+
106
+ More information needed
107
+
108
+ ## Training and evaluation data
109
+
110
+ More information needed
111
+
112
+ ## Training procedure
113
+
114
+ ### Training hyperparameters
115
+
116
+ The following hyperparameters were used during training:
117
+ - learning_rate: 0.0002
118
+ - train_batch_size: 1
119
+ - eval_batch_size: 1
120
+ - seed: 42
121
+ - distributed_type: multi-GPU
122
+ - num_devices: 8
123
+ - total_train_batch_size: 8
124
+ - total_eval_batch_size: 8
125
+ - optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
126
+ - lr_scheduler_type: cosine
127
+ - lr_scheduler_warmup_steps: 24
128
+ - num_epochs: 100
129
+
130
+ ### Training results
131
+
132
+ | Training Loss | Epoch | Step | Validation Loss |
133
+ |:-------------:|:-----:|:----:|:---------------:|
134
+ | No log | 0.025 | 1 | 0.9385 |
135
+ | 0.0292 | 1.0 | 40 | 0.0043 |
136
+ | 0.0148 | 2.0 | 80 | 0.0332 |
137
+ | 0.1015 | 3.0 | 120 | 0.0044 |
138
+ | 0.0002 | 4.0 | 160 | 0.0001 |
139
+ | 0.0 | 5.0 | 200 | 0.0000 |
140
+ | 0.0 | 6.0 | 240 | 0.0000 |
141
+ | 0.0 | 7.0 | 280 | 0.0000 |
142
+ | 0.0 | 8.0 | 320 | 0.0000 |
143
+ | 0.0 | 9.0 | 360 | 0.0000 |
144
+ | 0.0 | 10.0 | 400 | 0.0000 |
145
+ | 0.0 | 11.0 | 440 | 0.0000 |
146
+ | 0.0 | 12.0 | 480 | 0.0000 |
147
+ | 0.0 | 13.0 | 520 | 0.0000 |
148
+ | 0.0 | 14.0 | 560 | 0.0000 |
149
+ | 0.0 | 15.0 | 600 | 0.0000 |
150
+ | 0.0 | 16.0 | 640 | 0.0000 |
151
+ | 0.0 | 17.0 | 680 | 0.0000 |
152
+ | 0.0 | 18.0 | 720 | 0.0000 |
153
+ | 0.0 | 19.0 | 760 | 0.0000 |
154
+ | 0.0 | 20.0 | 800 | 0.0000 |
155
+ | 0.0 | 21.0 | 840 | 0.0000 |
156
+ | 0.0 | 22.0 | 880 | 0.0000 |
157
+ | 0.0 | 23.0 | 920 | 0.0000 |
158
+ | 0.0 | 24.0 | 960 | 0.0000 |
159
+ | 0.0 | 25.0 | 1000 | 0.0000 |
160
+ | 0.0 | 26.0 | 1040 | 0.0000 |
161
+ | 0.0 | 27.0 | 1080 | 0.0000 |
162
+ | 0.0 | 28.0 | 1120 | 0.0000 |
163
+ | 0.0 | 29.0 | 1160 | 0.0000 |
164
+ | 0.0 | 30.0 | 1200 | 0.0000 |
165
+ | 0.0 | 31.0 | 1240 | 0.0000 |
166
+ | 0.0 | 32.0 | 1280 | 0.0000 |
167
+ | 0.0 | 33.0 | 1320 | 0.0000 |
168
+ | 0.0 | 34.0 | 1360 | 0.0000 |
169
+ | 0.0 | 35.0 | 1400 | 0.0000 |
170
+ | 0.0 | 36.0 | 1440 | 0.0000 |
171
+ | 0.0 | 37.0 | 1480 | 0.0000 |
172
+ | 0.0 | 38.0 | 1520 | 0.0000 |
173
+ | 0.0 | 39.0 | 1560 | 0.0000 |
174
+ | 0.0 | 40.0 | 1600 | 0.0000 |
175
+ | 0.0 | 41.0 | 1640 | 0.0000 |
176
+ | 0.0 | 42.0 | 1680 | 0.0000 |
177
+ | 0.0 | 43.0 | 1720 | 0.0000 |
178
+ | 0.0 | 44.0 | 1760 | 0.0000 |
179
+ | 0.0 | 45.0 | 1800 | 0.0000 |
180
+ | 0.0 | 46.0 | 1840 | 0.0000 |
181
+ | 0.0 | 47.0 | 1880 | 0.0000 |
182
+ | 0.0 | 48.0 | 1920 | 0.0000 |
183
+ | 0.0 | 49.0 | 1960 | 0.0000 |
184
+ | 0.0 | 50.0 | 2000 | 0.0000 |
185
+ | 0.0 | 51.0 | 2040 | 0.0000 |
186
+ | 0.0 | 52.0 | 2080 | 0.0000 |
187
+ | 0.0 | 53.0 | 2120 | 0.0000 |
188
+ | 0.0 | 54.0 | 2160 | 0.0000 |
189
+ | 0.0 | 55.0 | 2200 | 0.0000 |
190
+ | 0.0 | 56.0 | 2240 | 0.0000 |
191
+ | 0.0 | 57.0 | 2280 | 0.0000 |
192
+ | 0.0 | 58.0 | 2320 | 0.0000 |
193
+ | 0.0 | 59.0 | 2360 | 0.0000 |
194
+ | 0.0 | 60.0 | 2400 | 0.0000 |
195
+ | 0.0 | 61.0 | 2440 | 0.0000 |
196
+ | 0.0 | 62.0 | 2480 | 0.0000 |
197
+ | 0.0 | 63.0 | 2520 | 0.0000 |
198
+ | 0.0 | 64.0 | 2560 | 0.0000 |
199
+ | 0.0 | 65.0 | 2600 | 0.0000 |
200
+ | 0.0 | 66.0 | 2640 | 0.0000 |
201
+ | 0.0 | 67.0 | 2680 | 0.0000 |
202
+ | 0.0 | 68.0 | 2720 | 0.0000 |
203
+ | 0.0 | 69.0 | 2760 | 0.0000 |
204
+ | 0.0 | 70.0 | 2800 | 0.0000 |
205
+ | 0.0 | 71.0 | 2840 | 0.0000 |
206
+ | 0.0 | 72.0 | 2880 | 0.0000 |
207
+ | 0.0 | 73.0 | 2920 | 0.0000 |
208
+ | 0.0 | 74.0 | 2960 | 0.0000 |
209
+ | 0.0 | 75.0 | 3000 | 0.0000 |
210
+ | 0.0 | 76.0 | 3040 | 0.0000 |
211
+ | 0.0 | 77.0 | 3080 | 0.0000 |
212
+ | 0.0 | 78.0 | 3120 | 0.0000 |
213
+ | 0.0 | 79.0 | 3160 | 0.0000 |
214
+ | 0.0 | 80.0 | 3200 | 0.0000 |
215
+ | 0.0 | 81.0 | 3240 | 0.0000 |
216
+ | 0.0 | 82.0 | 3280 | 0.0000 |
217
+ | 0.0 | 83.0 | 3320 | 0.0000 |
218
+ | 0.0 | 84.0 | 3360 | 0.0000 |
219
+ | 0.0 | 85.0 | 3400 | 0.0000 |
220
+ | 0.0 | 86.0 | 3440 | 0.0000 |
221
+ | 0.0 | 87.0 | 3480 | 0.0000 |
222
+ | 0.0 | 88.0 | 3520 | 0.0000 |
223
+ | 0.0 | 89.0 | 3560 | 0.0000 |
224
+ | 0.0 | 90.0 | 3600 | 0.0000 |
225
+ | 0.0 | 91.0 | 3640 | 0.0000 |
226
+ | 0.0 | 92.0 | 3680 | 0.0000 |
227
+ | 0.0 | 93.0 | 3720 | 0.0000 |
228
+ | 0.0 | 94.0 | 3760 | 0.0000 |
229
+ | 0.0 | 95.0 | 3800 | 0.0000 |
230
+ | 0.0 | 96.0 | 3840 | 0.0000 |
231
+ | 0.0 | 97.0 | 3880 | 0.0000 |
232
+ | 0.0 | 98.0 | 3920 | 0.0000 |
233
+ | 0.0 | 99.0 | 3960 | 0.0000 |
234
+ | 0.0 | 100.0 | 4000 | 0.0000 |
235
+
236
+
237
+ ### Framework versions
238
+
239
+ - PEFT 0.13.2
240
+ - Transformers 4.46.0
241
+ - Pytorch 2.5.0+cu124
242
+ - Datasets 3.0.1
243
  - Tokenizers 0.20.1