Hugging Face's logo

CharlesLi
/
mistral_default_lora

PEFT
TensorBoard
Safetensors
mistral
alignment-handbook
trl
sft
Generated from Trainer
Model card Files
xet
Metrics Community
CharlesLi commited on
Commit
680eebf
·
verified ·
1 Parent(s): 20ee032

Model save

Browse files
Files changed (4) hide show
  1. README.md +81 -0
  2. all_results.json +9 -0
  3. train_results.json +9 -0
  4. trainer_state.json +190 -0
README.md ADDED
@@ -0,0 +1,81 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: mistralai/Mistral-7B-Instruct-v0.1
3
+ datasets:
4
+ - generator
5
+ library_name: peft
6
+ license: apache-2.0
7
+ tags:
8
+ - trl
9
+ - sft
10
+ - generated_from_trainer
11
+ model-index:
12
+ - name: mistral_default_lora
13
+ results: []
14
+ ---
15
+
16
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
17
+ should probably proofread and complete it, then remove this comment. -->
18
+
19
+ # mistral_default_lora
20
+
21
+ This model is a fine-tuned version of [mistralai/Mistral-7B-Instruct-v0.1](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.1) on the generator dataset.
22
+ It achieves the following results on the evaluation set:
23
+ - Loss: 1.2007
24
+
25
+ ## Model description
26
+
27
+ More information needed
28
+
29
+ ## Intended uses & limitations
30
+
31
+ More information needed
32
+
33
+ ## Training and evaluation data
34
+
35
+ More information needed
36
+
37
+ ## Training procedure
38
+
39
+ ### Training hyperparameters
40
+
41
+ The following hyperparameters were used during training:
42
+ - learning_rate: 0.0002
43
+ - train_batch_size: 4
44
+ - eval_batch_size: 4
45
+ - seed: 42
46
+ - distributed_type: multi-GPU
47
+ - gradient_accumulation_steps: 2
48
+ - total_train_batch_size: 8
49
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
50
+ - lr_scheduler_type: cosine
51
+ - lr_scheduler_warmup_ratio: 0.1
52
+ - num_epochs: 15
53
+
54
+ ### Training results
55
+
56
+ | Training Loss | Epoch | Step | Validation Loss |
57
+ |:-------------:|:-----:|:----:|:---------------:|
58
+ | 2.054 | 1.0 | 1 | 2.0482 |
59
+ | 2.054 | 2.0 | 2 | 1.7447 |
60
+ | 2.054 | 3.0 | 3 | 1.4612 |
61
+ | 2.054 | 4.0 | 4 | 1.3664 |
62
+ | 1.636 | 5.0 | 5 | 1.3182 |
63
+ | 1.636 | 6.0 | 6 | 1.2509 |
64
+ | 1.636 | 7.0 | 7 | 1.2262 |
65
+ | 1.636 | 8.0 | 8 | 1.2207 |
66
+ | 1.636 | 9.0 | 9 | 1.2038 |
67
+ | 0.8947 | 10.0 | 10 | 1.1923 |
68
+ | 0.8947 | 11.0 | 11 | 1.1926 |
69
+ | 0.8947 | 12.0 | 12 | 1.1967 |
70
+ | 0.8947 | 13.0 | 13 | 1.1993 |
71
+ | 0.8947 | 14.0 | 14 | 1.2008 |
72
+ | 0.5319 | 15.0 | 15 | 1.2007 |
73
+
74
+
75
+ ### Framework versions
76
+
77
+ - PEFT 0.12.0
78
+ - Transformers 4.44.2
79
+ - Pytorch 2.4.1+cu121
80
+ - Datasets 3.0.0
81
+ - Tokenizers 0.19.1
all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 15.0,
3
+ "total_flos": 2787467329536.0,
4
+ "train_loss": 1.0487482070922851,
5
+ "train_runtime": 88.2648,
6
+ "train_samples": 100,
7
+ "train_samples_per_second": 1.02,
8
+ "train_steps_per_second": 0.17
9
+ }
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 15.0,
3
+ "total_flos": 2787467329536.0,
4
+ "train_loss": 1.0487482070922851,
5
+ "train_runtime": 88.2648,
6
+ "train_samples": 100,
7
+ "train_samples_per_second": 1.02,
8
+ "train_steps_per_second": 0.17
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,190 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 15.0,
5
+ "eval_steps": 500,
6
+ "global_step": 15,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 1.0,
13
+ "grad_norm": 2.775644201582239,
14
+ "learning_rate": 0.0001,
15
+ "loss": 2.054,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 1.0,
20
+ "eval_loss": 2.0482397079467773,
21
+ "eval_runtime": 0.4302,
22
+ "eval_samples_per_second": 2.324,
23
+ "eval_steps_per_second": 2.324,
24
+ "step": 1
25
+ },
26
+ {
27
+ "epoch": 2.0,
28
+ "eval_loss": 1.7446707487106323,
29
+ "eval_runtime": 1.5149,
30
+ "eval_samples_per_second": 0.66,
31
+ "eval_steps_per_second": 0.66,
32
+ "step": 2
33
+ },
34
+ {
35
+ "epoch": 3.0,
36
+ "eval_loss": 1.4611855745315552,
37
+ "eval_runtime": 0.37,
38
+ "eval_samples_per_second": 2.703,
39
+ "eval_steps_per_second": 2.703,
40
+ "step": 3
41
+ },
42
+ {
43
+ "epoch": 4.0,
44
+ "eval_loss": 1.3664478063583374,
45
+ "eval_runtime": 0.3722,
46
+ "eval_samples_per_second": 2.686,
47
+ "eval_steps_per_second": 2.686,
48
+ "step": 4
49
+ },
50
+ {
51
+ "epoch": 5.0,
52
+ "grad_norm": 0.7397193908660463,
53
+ "learning_rate": 0.00017485107481711012,
54
+ "loss": 1.636,
55
+ "step": 5
56
+ },
57
+ {
58
+ "epoch": 5.0,
59
+ "eval_loss": 1.3181551694869995,
60
+ "eval_runtime": 0.3714,
61
+ "eval_samples_per_second": 2.693,
62
+ "eval_steps_per_second": 2.693,
63
+ "step": 5
64
+ },
65
+ {
66
+ "epoch": 6.0,
67
+ "eval_loss": 1.2508817911148071,
68
+ "eval_runtime": 0.3699,
69
+ "eval_samples_per_second": 2.704,
70
+ "eval_steps_per_second": 2.704,
71
+ "step": 6
72
+ },
73
+ {
74
+ "epoch": 7.0,
75
+ "eval_loss": 1.2261731624603271,
76
+ "eval_runtime": 0.3682,
77
+ "eval_samples_per_second": 2.716,
78
+ "eval_steps_per_second": 2.716,
79
+ "step": 7
80
+ },
81
+ {
82
+ "epoch": 8.0,
83
+ "eval_loss": 1.2207050323486328,
84
+ "eval_runtime": 0.3695,
85
+ "eval_samples_per_second": 2.706,
86
+ "eval_steps_per_second": 2.706,
87
+ "step": 8
88
+ },
89
+ {
90
+ "epoch": 9.0,
91
+ "eval_loss": 1.2037922143936157,
92
+ "eval_runtime": 0.3696,
93
+ "eval_samples_per_second": 2.706,
94
+ "eval_steps_per_second": 2.706,
95
+ "step": 9
96
+ },
97
+ {
98
+ "epoch": 10.0,
99
+ "grad_norm": 0.7005825834442886,
100
+ "learning_rate": 6.453951129574644e-05,
101
+ "loss": 0.8947,
102
+ "step": 10
103
+ },
104
+ {
105
+ "epoch": 10.0,
106
+ "eval_loss": 1.1922813653945923,
107
+ "eval_runtime": 0.37,
108
+ "eval_samples_per_second": 2.703,
109
+ "eval_steps_per_second": 2.703,
110
+ "step": 10
111
+ },
112
+ {
113
+ "epoch": 11.0,
114
+ "eval_loss": 1.1926474571228027,
115
+ "eval_runtime": 0.3699,
116
+ "eval_samples_per_second": 2.703,
117
+ "eval_steps_per_second": 2.703,
118
+ "step": 11
119
+ },
120
+ {
121
+ "epoch": 12.0,
122
+ "eval_loss": 1.1966757774353027,
123
+ "eval_runtime": 0.3703,
124
+ "eval_samples_per_second": 2.7,
125
+ "eval_steps_per_second": 2.7,
126
+ "step": 12
127
+ },
128
+ {
129
+ "epoch": 13.0,
130
+ "eval_loss": 1.1992850303649902,
131
+ "eval_runtime": 0.3694,
132
+ "eval_samples_per_second": 2.707,
133
+ "eval_steps_per_second": 2.707,
134
+ "step": 13
135
+ },
136
+ {
137
+ "epoch": 14.0,
138
+ "eval_loss": 1.200769305229187,
139
+ "eval_runtime": 0.3692,
140
+ "eval_samples_per_second": 2.709,
141
+ "eval_steps_per_second": 2.709,
142
+ "step": 14
143
+ },
144
+ {
145
+ "epoch": 15.0,
146
+ "grad_norm": 0.5467216766155176,
147
+ "learning_rate": 0.0,
148
+ "loss": 0.5319,
149
+ "step": 15
150
+ },
151
+ {
152
+ "epoch": 15.0,
153
+ "eval_loss": 1.200675368309021,
154
+ "eval_runtime": 0.3831,
155
+ "eval_samples_per_second": 2.61,
156
+ "eval_steps_per_second": 2.61,
157
+ "step": 15
158
+ },
159
+ {
160
+ "epoch": 15.0,
161
+ "step": 15,
162
+ "total_flos": 2787467329536.0,
163
+ "train_loss": 1.0487482070922851,
164
+ "train_runtime": 88.2648,
165
+ "train_samples_per_second": 1.02,
166
+ "train_steps_per_second": 0.17
167
+ }
168
+ ],
169
+ "logging_steps": 5,
170
+ "max_steps": 15,
171
+ "num_input_tokens_seen": 0,
172
+ "num_train_epochs": 15,
173
+ "save_steps": 1000,
174
+ "stateful_callbacks": {
175
+ "TrainerControl": {
176
+ "args": {
177
+ "should_epoch_stop": false,
178
+ "should_evaluate": false,
179
+ "should_log": false,
180
+ "should_save": true,
181
+ "should_training_stop": true
182
+ },
183
+ "attributes": {}
184
+ }
185
+ },
186
+ "total_flos": 2787467329536.0,
187
+ "train_batch_size": 4,
188
+ "trial_name": null,
189
+ "trial_params": null
190
+ }