Hugging Face's logo

bigheiniuJ
/
zephyr-7b-feedforward-full

Text Generation
Transformers
TensorBoard
Safetensors
mistral
Generated from Trainer
trl
sft
conversational
text-generation-inference
Model card Files Metrics Community
bigheiniuJ commited on
Commit
5b85d18
·
verified ·
1 Parent(s): e8743f4

Model save

Browse files
Files changed (9) hide show
  1. README.md +57 -0
  2. all_results.json +9 -0
  3. generation_config.json +6 -0
  4. model-00001-of-00003.safetensors +1 -1
  5. model-00002-of-00003.safetensors +1 -1
  6. model-00003-of-00003.safetensors +1 -1
  7. runs/Dec07_20-58-59_ip-10-0-8-86.ec2.internal/events.out.tfevents.1733605670.ip-10-0-8-86.ec2.internal.35913.0 +2 -2
  8. train_results.json +9 -0
  9. trainer_state.json +283 -0
README.md ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: alignment-handbook/zephyr-7b-sft-full
3
+ library_name: transformers
4
+ model_name: zephyr-7b-feedforward-full
5
+ tags:
6
+ - generated_from_trainer
7
+ - trl
8
+ - sft
9
+ licence: license
10
+ ---
11
+
12
+ # Model Card for zephyr-7b-feedforward-full
13
+
14
+ This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full).
15
+ It has been trained using [TRL](https://github.com/huggingface/trl).
16
+
17
+ ## Quick start
18
+
19
+ ```python
20
+ from transformers import pipeline
21
+
22
+ question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
23
+ generator = pipeline("text-generation", model="bigheiniuJ/zephyr-7b-feedforward-full", device="cuda")
24
+ output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
25
+ print(output["generated_text"])
26
+ ```
27
+
28
+ ## Training procedure
29
+
30
+
31
+
32
+ This model was trained with SFT.
33
+
34
+ ### Framework versions
35
+
36
+ - TRL: 0.12.1
37
+ - Transformers: 4.46.3
38
+ - Pytorch: 2.5.1
39
+ - Datasets: 2.21.0
40
+ - Tokenizers: 0.20.3
41
+
42
+ ## Citations
43
+
44
+
45
+
46
+ Cite TRL as:
47
+
48
+ ```bibtex
49
+ @misc{vonwerra2022trl,
50
+ title = {{TRL: Transformer Reinforcement Learning}},
51
+ author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
52
+ year = 2020,
53
+ journal = {GitHub repository},
54
+ publisher = {GitHub},
55
+ howpublished = {\url{https://github.com/huggingface/trl}}
56
+ }
57
+ ```
all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "total_flos": 126255932375040.0,
4
+ "train_loss": 1.1442423174712832,
5
+ "train_runtime": 3706.3256,
6
+ "train_samples": 122269,
7
+ "train_samples_per_second": 10.417,
8
+ "train_steps_per_second": 0.081
9
+ }
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": "4.46.3"
6
+ }
model-00001-of-00003.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:42bf36406a007e1f93cc3a4da4be2fb40757e3c3852cdf6abd4f87228afeeb3d
3
  size 4943162336
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:293a9612dc54349939fee1e9b1afd1fe6e312b819454fe3fa9b88bd081eeb5c6
3
  size 4943162336
model-00002-of-00003.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3c48442977fd4901f7db0a37454e225e35abbf47f5bfc8aeeeac98497ddb71c2
3
  size 4999819336
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0b0231809c003a252782856542698bbeb49faff867011541507ac690a72cf9e8
3
  size 4999819336
model-00003-of-00003.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:68f84b0b70e90a7b73a055ae43757bb47cfc4d1da641996c40a30ae5b2c5d100
3
  size 4540516344
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fca940be5662f7ec362320fcf7471c423e9500fc4d37561274ca504d6c82a6e1
3
  size 4540516344
runs/Dec07_20-58-59_ip-10-0-8-86.ec2.internal/events.out.tfevents.1733605670.ip-10-0-8-86.ec2.internal.35913.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:7f7d4eed1b46d9fc6e640b9b6f0fd86b801451394490525a2dc1ce756e7c3938
3
- size 12967
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:68de8b9e39a61c05339e42876beb38c5f57ff363ed6278faaa45a36da72c0584
3
+ size 13321
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "total_flos": 126255932375040.0,
4
+ "train_loss": 1.1442423174712832,
5
+ "train_runtime": 3706.3256,
6
+ "train_samples": 122269,
7
+ "train_samples_per_second": 10.417,
8
+ "train_steps_per_second": 0.081
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,283 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.0,
5
+ "eval_steps": 100,
6
+ "global_step": 302,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0033112582781456954,
13
+ "grad_norm": 8.547256141414913,
14
+ "learning_rate": 1.6129032258064514e-08,
15
+ "loss": 1.3889,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 0.033112582781456956,
20
+ "grad_norm": 8.048209966870028,
21
+ "learning_rate": 1.6129032258064515e-07,
22
+ "loss": 1.4125,
23
+ "step": 10
24
+ },
25
+ {
26
+ "epoch": 0.06622516556291391,
27
+ "grad_norm": 4.441313871275738,
28
+ "learning_rate": 3.225806451612903e-07,
29
+ "loss": 1.3764,
30
+ "step": 20
31
+ },
32
+ {
33
+ "epoch": 0.09933774834437085,
34
+ "grad_norm": 2.792727195498978,
35
+ "learning_rate": 4.838709677419355e-07,
36
+ "loss": 1.297,
37
+ "step": 30
38
+ },
39
+ {
40
+ "epoch": 0.13245033112582782,
41
+ "grad_norm": 2.440195994000396,
42
+ "learning_rate": 4.98640553244975e-07,
43
+ "loss": 1.2474,
44
+ "step": 40
45
+ },
46
+ {
47
+ "epoch": 0.16556291390728478,
48
+ "grad_norm": 1.7402831463764055,
49
+ "learning_rate": 4.939602093872018e-07,
50
+ "loss": 1.2119,
51
+ "step": 50
52
+ },
53
+ {
54
+ "epoch": 0.1986754966887417,
55
+ "grad_norm": 1.8044881066728908,
56
+ "learning_rate": 4.860049979829031e-07,
57
+ "loss": 1.1844,
58
+ "step": 60
59
+ },
60
+ {
61
+ "epoch": 0.23178807947019867,
62
+ "grad_norm": 2.4217209978341887,
63
+ "learning_rate": 4.7488170801918125e-07,
64
+ "loss": 1.1695,
65
+ "step": 70
66
+ },
67
+ {
68
+ "epoch": 0.26490066225165565,
69
+ "grad_norm": 1.6559949763449162,
70
+ "learning_rate": 4.6073965606452205e-07,
71
+ "loss": 1.1473,
72
+ "step": 80
73
+ },
74
+ {
75
+ "epoch": 0.2980132450331126,
76
+ "grad_norm": 1.6318360522717992,
77
+ "learning_rate": 4.437686818764167e-07,
78
+ "loss": 1.1446,
79
+ "step": 90
80
+ },
81
+ {
82
+ "epoch": 0.33112582781456956,
83
+ "grad_norm": 1.5824289985736384,
84
+ "learning_rate": 4.241966000347147e-07,
85
+ "loss": 1.1247,
86
+ "step": 100
87
+ },
88
+ {
89
+ "epoch": 0.33112582781456956,
90
+ "eval_loss": 1.1251499652862549,
91
+ "eval_runtime": 28.1687,
92
+ "eval_samples_per_second": 44.375,
93
+ "eval_steps_per_second": 0.71,
94
+ "step": 100
95
+ },
96
+ {
97
+ "epoch": 0.36423841059602646,
98
+ "grad_norm": 1.6700627307800546,
99
+ "learning_rate": 4.0228614180941443e-07,
100
+ "loss": 1.127,
101
+ "step": 110
102
+ },
103
+ {
104
+ "epoch": 0.3973509933774834,
105
+ "grad_norm": 1.6527451316364625,
106
+ "learning_rate": 3.7833142831457683e-07,
107
+ "loss": 1.1282,
108
+ "step": 120
109
+ },
110
+ {
111
+ "epoch": 0.4304635761589404,
112
+ "grad_norm": 1.7456316161527055,
113
+ "learning_rate": 3.526540222919605e-07,
114
+ "loss": 1.1212,
115
+ "step": 130
116
+ },
117
+ {
118
+ "epoch": 0.46357615894039733,
119
+ "grad_norm": 1.6991459409500143,
120
+ "learning_rate": 3.255986115243545e-07,
121
+ "loss": 1.1275,
122
+ "step": 140
123
+ },
124
+ {
125
+ "epoch": 0.4966887417218543,
126
+ "grad_norm": 1.7534210030423583,
127
+ "learning_rate": 2.97528381823508e-07,
128
+ "loss": 1.1049,
129
+ "step": 150
130
+ },
131
+ {
132
+ "epoch": 0.5298013245033113,
133
+ "grad_norm": 1.7140941720199787,
134
+ "learning_rate": 2.68820141704636e-07,
135
+ "loss": 1.1042,
136
+ "step": 160
137
+ },
138
+ {
139
+ "epoch": 0.5629139072847682,
140
+ "grad_norm": 1.7525269414058429,
141
+ "learning_rate": 2.398592641927878e-07,
142
+ "loss": 1.0951,
143
+ "step": 170
144
+ },
145
+ {
146
+ "epoch": 0.5960264900662252,
147
+ "grad_norm": 1.6093519147740063,
148
+ "learning_rate": 2.1103451366114007e-07,
149
+ "loss": 1.11,
150
+ "step": 180
151
+ },
152
+ {
153
+ "epoch": 0.6291390728476821,
154
+ "grad_norm": 1.7141231135481751,
155
+ "learning_rate": 1.827328271445881e-07,
156
+ "loss": 1.1127,
157
+ "step": 190
158
+ },
159
+ {
160
+ "epoch": 0.6622516556291391,
161
+ "grad_norm": 1.6936730364698183,
162
+ "learning_rate": 1.5533412018312042e-07,
163
+ "loss": 1.1054,
164
+ "step": 200
165
+ },
166
+ {
167
+ "epoch": 0.6622516556291391,
168
+ "eval_loss": 1.0929443836212158,
169
+ "eval_runtime": 26.7241,
170
+ "eval_samples_per_second": 46.774,
171
+ "eval_steps_per_second": 0.748,
172
+ "step": 200
173
+ },
174
+ {
175
+ "epoch": 0.695364238410596,
176
+ "grad_norm": 1.6640304962792698,
177
+ "learning_rate": 1.2920618692017632e-07,
178
+ "loss": 1.0931,
179
+ "step": 210
180
+ },
181
+ {
182
+ "epoch": 0.7284768211920529,
183
+ "grad_norm": 1.6853199575869429,
184
+ "learning_rate": 1.0469976291593402e-07,
185
+ "loss": 1.0786,
186
+ "step": 220
187
+ },
188
+ {
189
+ "epoch": 0.7615894039735099,
190
+ "grad_norm": 1.6787446835035877,
191
+ "learning_rate": 8.21438169512223e-08,
192
+ "loss": 1.0944,
193
+ "step": 230
194
+ },
195
+ {
196
+ "epoch": 0.7947019867549668,
197
+ "grad_norm": 1.7361953819881613,
198
+ "learning_rate": 6.184113502383415e-08,
199
+ "loss": 1.0895,
200
+ "step": 240
201
+ },
202
+ {
203
+ "epoch": 0.8278145695364238,
204
+ "grad_norm": 1.6915106308947352,
205
+ "learning_rate": 4.406425581668988e-08,
206
+ "loss": 1.0921,
207
+ "step": 250
208
+ },
209
+ {
210
+ "epoch": 0.8609271523178808,
211
+ "grad_norm": 1.6890053704917347,
212
+ "learning_rate": 2.905181219922448e-08,
213
+ "loss": 1.0888,
214
+ "step": 260
215
+ },
216
+ {
217
+ "epoch": 0.8940397350993378,
218
+ "grad_norm": 1.7278129612109874,
219
+ "learning_rate": 1.700532787286593e-08,
220
+ "loss": 1.0897,
221
+ "step": 270
222
+ },
223
+ {
224
+ "epoch": 0.9271523178807947,
225
+ "grad_norm": 1.7209342188458445,
226
+ "learning_rate": 8.086512161725873e-09,
227
+ "loss": 1.0865,
228
+ "step": 280
229
+ },
230
+ {
231
+ "epoch": 0.9602649006622517,
232
+ "grad_norm": 1.68192345667153,
233
+ "learning_rate": 2.4150892626280394e-09,
234
+ "loss": 1.0884,
235
+ "step": 290
236
+ },
237
+ {
238
+ "epoch": 0.9933774834437086,
239
+ "grad_norm": 1.6887236258880367,
240
+ "learning_rate": 6.719109414038416e-11,
241
+ "loss": 1.0915,
242
+ "step": 300
243
+ },
244
+ {
245
+ "epoch": 0.9933774834437086,
246
+ "eval_loss": 1.087569236755371,
247
+ "eval_runtime": 27.0413,
248
+ "eval_samples_per_second": 46.226,
249
+ "eval_steps_per_second": 0.74,
250
+ "step": 300
251
+ },
252
+ {
253
+ "epoch": 1.0,
254
+ "step": 302,
255
+ "total_flos": 126255932375040.0,
256
+ "train_loss": 1.1442423174712832,
257
+ "train_runtime": 3706.3256,
258
+ "train_samples_per_second": 10.417,
259
+ "train_steps_per_second": 0.081
260
+ }
261
+ ],
262
+ "logging_steps": 10,
263
+ "max_steps": 302,
264
+ "num_input_tokens_seen": 0,
265
+ "num_train_epochs": 1,
266
+ "save_steps": 100,
267
+ "stateful_callbacks": {
268
+ "TrainerControl": {
269
+ "args": {
270
+ "should_epoch_stop": false,
271
+ "should_evaluate": false,
272
+ "should_log": false,
273
+ "should_save": true,
274
+ "should_training_stop": true
275
+ },
276
+ "attributes": {}
277
+ }
278
+ },
279
+ "total_flos": 126255932375040.0,
280
+ "train_batch_size": 8,
281
+ "trial_name": null,
282
+ "trial_params": null
283
+ }