End of training
Browse files- README.md +2 -1
- all_results.json +12 -0
- eval_results.json +7 -0
- train_results.json +8 -0
- trainer_state.json +668 -0
- training_eval_loss.png +0 -0
- training_loss.png +0 -0
README.md
CHANGED
@@ -4,6 +4,7 @@ license: llama3.1
|
|
4 |
base_model: meta-llama/Meta-Llama-3.1-8B
|
5 |
tags:
|
6 |
- llama-factory
|
|
|
7 |
- generated_from_trainer
|
8 |
model-index:
|
9 |
- name: llama3-1_8b_oh_dcft_v1_wo_metamath
|
@@ -15,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
15 |
|
16 |
# llama3-1_8b_oh_dcft_v1_wo_metamath
|
17 |
|
18 |
-
This model is a fine-tuned version of [meta-llama/Meta-Llama-3.1-8B](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B) on
|
19 |
It achieves the following results on the evaluation set:
|
20 |
- Loss: 0.5936
|
21 |
|
|
|
4 |
base_model: meta-llama/Meta-Llama-3.1-8B
|
5 |
tags:
|
6 |
- llama-factory
|
7 |
+
- full
|
8 |
- generated_from_trainer
|
9 |
model-index:
|
10 |
- name: llama3-1_8b_oh_dcft_v1_wo_metamath
|
|
|
16 |
|
17 |
# llama3-1_8b_oh_dcft_v1_wo_metamath
|
18 |
|
19 |
+
This model is a fine-tuned version of [meta-llama/Meta-Llama-3.1-8B](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B) on the mlfoundations-dev/OH_DCFT_v1_wo_metamath dataset.
|
20 |
It achieves the following results on the evaluation set:
|
21 |
- Loss: 0.5936
|
22 |
|
all_results.json
ADDED
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 3.0,
|
3 |
+
"eval_loss": 0.5936245918273926,
|
4 |
+
"eval_runtime": 26.9443,
|
5 |
+
"eval_samples_per_second": 288.929,
|
6 |
+
"eval_steps_per_second": 1.151,
|
7 |
+
"total_flos": 4.093705665380352e+19,
|
8 |
+
"train_loss": 0.5755822655788893,
|
9 |
+
"train_runtime": 6372.9887,
|
10 |
+
"train_samples_per_second": 69.622,
|
11 |
+
"train_steps_per_second": 0.136
|
12 |
+
}
|
eval_results.json
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 3.0,
|
3 |
+
"eval_loss": 0.5936245918273926,
|
4 |
+
"eval_runtime": 26.9443,
|
5 |
+
"eval_samples_per_second": 288.929,
|
6 |
+
"eval_steps_per_second": 1.151
|
7 |
+
}
|
train_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 3.0,
|
3 |
+
"total_flos": 4.093705665380352e+19,
|
4 |
+
"train_loss": 0.5755822655788893,
|
5 |
+
"train_runtime": 6372.9887,
|
6 |
+
"train_samples_per_second": 69.622,
|
7 |
+
"train_steps_per_second": 0.136
|
8 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,668 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 3.0,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 867,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.03460207612456748,
|
13 |
+
"grad_norm": 5.743846893310547,
|
14 |
+
"learning_rate": 5e-06,
|
15 |
+
"loss": 0.9289,
|
16 |
+
"step": 10
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 0.06920415224913495,
|
20 |
+
"grad_norm": 5.331993103027344,
|
21 |
+
"learning_rate": 5e-06,
|
22 |
+
"loss": 0.8126,
|
23 |
+
"step": 20
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.10380622837370242,
|
27 |
+
"grad_norm": 5.425512790679932,
|
28 |
+
"learning_rate": 5e-06,
|
29 |
+
"loss": 0.7742,
|
30 |
+
"step": 30
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"epoch": 0.1384083044982699,
|
34 |
+
"grad_norm": 1.6510249376296997,
|
35 |
+
"learning_rate": 5e-06,
|
36 |
+
"loss": 0.747,
|
37 |
+
"step": 40
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.17301038062283736,
|
41 |
+
"grad_norm": 1.0029033422470093,
|
42 |
+
"learning_rate": 5e-06,
|
43 |
+
"loss": 0.7119,
|
44 |
+
"step": 50
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.20761245674740483,
|
48 |
+
"grad_norm": 1.1877987384796143,
|
49 |
+
"learning_rate": 5e-06,
|
50 |
+
"loss": 0.688,
|
51 |
+
"step": 60
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.2422145328719723,
|
55 |
+
"grad_norm": 1.0992727279663086,
|
56 |
+
"learning_rate": 5e-06,
|
57 |
+
"loss": 0.6709,
|
58 |
+
"step": 70
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"epoch": 0.2768166089965398,
|
62 |
+
"grad_norm": 1.0378435850143433,
|
63 |
+
"learning_rate": 5e-06,
|
64 |
+
"loss": 0.6633,
|
65 |
+
"step": 80
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.31141868512110726,
|
69 |
+
"grad_norm": 0.7307208776473999,
|
70 |
+
"learning_rate": 5e-06,
|
71 |
+
"loss": 0.659,
|
72 |
+
"step": 90
|
73 |
+
},
|
74 |
+
{
|
75 |
+
"epoch": 0.3460207612456747,
|
76 |
+
"grad_norm": 0.8541375994682312,
|
77 |
+
"learning_rate": 5e-06,
|
78 |
+
"loss": 0.6368,
|
79 |
+
"step": 100
|
80 |
+
},
|
81 |
+
{
|
82 |
+
"epoch": 0.3806228373702422,
|
83 |
+
"grad_norm": 0.9289709329605103,
|
84 |
+
"learning_rate": 5e-06,
|
85 |
+
"loss": 0.642,
|
86 |
+
"step": 110
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 0.41522491349480967,
|
90 |
+
"grad_norm": 0.6918668746948242,
|
91 |
+
"learning_rate": 5e-06,
|
92 |
+
"loss": 0.6332,
|
93 |
+
"step": 120
|
94 |
+
},
|
95 |
+
{
|
96 |
+
"epoch": 0.44982698961937717,
|
97 |
+
"grad_norm": 0.7631871700286865,
|
98 |
+
"learning_rate": 5e-06,
|
99 |
+
"loss": 0.6317,
|
100 |
+
"step": 130
|
101 |
+
},
|
102 |
+
{
|
103 |
+
"epoch": 0.4844290657439446,
|
104 |
+
"grad_norm": 0.6295446157455444,
|
105 |
+
"learning_rate": 5e-06,
|
106 |
+
"loss": 0.6308,
|
107 |
+
"step": 140
|
108 |
+
},
|
109 |
+
{
|
110 |
+
"epoch": 0.5190311418685121,
|
111 |
+
"grad_norm": 0.635179340839386,
|
112 |
+
"learning_rate": 5e-06,
|
113 |
+
"loss": 0.6297,
|
114 |
+
"step": 150
|
115 |
+
},
|
116 |
+
{
|
117 |
+
"epoch": 0.5536332179930796,
|
118 |
+
"grad_norm": 0.7158755660057068,
|
119 |
+
"learning_rate": 5e-06,
|
120 |
+
"loss": 0.6251,
|
121 |
+
"step": 160
|
122 |
+
},
|
123 |
+
{
|
124 |
+
"epoch": 0.5882352941176471,
|
125 |
+
"grad_norm": 0.8344781994819641,
|
126 |
+
"learning_rate": 5e-06,
|
127 |
+
"loss": 0.609,
|
128 |
+
"step": 170
|
129 |
+
},
|
130 |
+
{
|
131 |
+
"epoch": 0.6228373702422145,
|
132 |
+
"grad_norm": 0.7887862324714661,
|
133 |
+
"learning_rate": 5e-06,
|
134 |
+
"loss": 0.6186,
|
135 |
+
"step": 180
|
136 |
+
},
|
137 |
+
{
|
138 |
+
"epoch": 0.657439446366782,
|
139 |
+
"grad_norm": 0.7533692717552185,
|
140 |
+
"learning_rate": 5e-06,
|
141 |
+
"loss": 0.6201,
|
142 |
+
"step": 190
|
143 |
+
},
|
144 |
+
{
|
145 |
+
"epoch": 0.6920415224913494,
|
146 |
+
"grad_norm": 0.7703404426574707,
|
147 |
+
"learning_rate": 5e-06,
|
148 |
+
"loss": 0.6176,
|
149 |
+
"step": 200
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 0.726643598615917,
|
153 |
+
"grad_norm": 0.6667165160179138,
|
154 |
+
"learning_rate": 5e-06,
|
155 |
+
"loss": 0.612,
|
156 |
+
"step": 210
|
157 |
+
},
|
158 |
+
{
|
159 |
+
"epoch": 0.7612456747404844,
|
160 |
+
"grad_norm": 0.7431259155273438,
|
161 |
+
"learning_rate": 5e-06,
|
162 |
+
"loss": 0.6069,
|
163 |
+
"step": 220
|
164 |
+
},
|
165 |
+
{
|
166 |
+
"epoch": 0.7958477508650519,
|
167 |
+
"grad_norm": 0.6150693893432617,
|
168 |
+
"learning_rate": 5e-06,
|
169 |
+
"loss": 0.6055,
|
170 |
+
"step": 230
|
171 |
+
},
|
172 |
+
{
|
173 |
+
"epoch": 0.8304498269896193,
|
174 |
+
"grad_norm": 1.0219345092773438,
|
175 |
+
"learning_rate": 5e-06,
|
176 |
+
"loss": 0.6068,
|
177 |
+
"step": 240
|
178 |
+
},
|
179 |
+
{
|
180 |
+
"epoch": 0.8650519031141869,
|
181 |
+
"grad_norm": 0.5619581341743469,
|
182 |
+
"learning_rate": 5e-06,
|
183 |
+
"loss": 0.6026,
|
184 |
+
"step": 250
|
185 |
+
},
|
186 |
+
{
|
187 |
+
"epoch": 0.8996539792387543,
|
188 |
+
"grad_norm": 0.5808368921279907,
|
189 |
+
"learning_rate": 5e-06,
|
190 |
+
"loss": 0.6077,
|
191 |
+
"step": 260
|
192 |
+
},
|
193 |
+
{
|
194 |
+
"epoch": 0.9342560553633218,
|
195 |
+
"grad_norm": 0.6060568690299988,
|
196 |
+
"learning_rate": 5e-06,
|
197 |
+
"loss": 0.6031,
|
198 |
+
"step": 270
|
199 |
+
},
|
200 |
+
{
|
201 |
+
"epoch": 0.9688581314878892,
|
202 |
+
"grad_norm": 0.515640139579773,
|
203 |
+
"learning_rate": 5e-06,
|
204 |
+
"loss": 0.6051,
|
205 |
+
"step": 280
|
206 |
+
},
|
207 |
+
{
|
208 |
+
"epoch": 1.0,
|
209 |
+
"eval_loss": 0.6054237484931946,
|
210 |
+
"eval_runtime": 26.5711,
|
211 |
+
"eval_samples_per_second": 292.988,
|
212 |
+
"eval_steps_per_second": 1.167,
|
213 |
+
"step": 289
|
214 |
+
},
|
215 |
+
{
|
216 |
+
"epoch": 1.0034602076124568,
|
217 |
+
"grad_norm": 1.3408666849136353,
|
218 |
+
"learning_rate": 5e-06,
|
219 |
+
"loss": 0.5952,
|
220 |
+
"step": 290
|
221 |
+
},
|
222 |
+
{
|
223 |
+
"epoch": 1.0380622837370241,
|
224 |
+
"grad_norm": 1.0180834531784058,
|
225 |
+
"learning_rate": 5e-06,
|
226 |
+
"loss": 0.558,
|
227 |
+
"step": 300
|
228 |
+
},
|
229 |
+
{
|
230 |
+
"epoch": 1.0726643598615917,
|
231 |
+
"grad_norm": 0.6628531217575073,
|
232 |
+
"learning_rate": 5e-06,
|
233 |
+
"loss": 0.5671,
|
234 |
+
"step": 310
|
235 |
+
},
|
236 |
+
{
|
237 |
+
"epoch": 1.1072664359861593,
|
238 |
+
"grad_norm": 0.6421555876731873,
|
239 |
+
"learning_rate": 5e-06,
|
240 |
+
"loss": 0.5575,
|
241 |
+
"step": 320
|
242 |
+
},
|
243 |
+
{
|
244 |
+
"epoch": 1.1418685121107266,
|
245 |
+
"grad_norm": 0.7754614353179932,
|
246 |
+
"learning_rate": 5e-06,
|
247 |
+
"loss": 0.5663,
|
248 |
+
"step": 330
|
249 |
+
},
|
250 |
+
{
|
251 |
+
"epoch": 1.1764705882352942,
|
252 |
+
"grad_norm": 0.7297713160514832,
|
253 |
+
"learning_rate": 5e-06,
|
254 |
+
"loss": 0.5553,
|
255 |
+
"step": 340
|
256 |
+
},
|
257 |
+
{
|
258 |
+
"epoch": 1.2110726643598615,
|
259 |
+
"grad_norm": 0.7665261030197144,
|
260 |
+
"learning_rate": 5e-06,
|
261 |
+
"loss": 0.5524,
|
262 |
+
"step": 350
|
263 |
+
},
|
264 |
+
{
|
265 |
+
"epoch": 1.245674740484429,
|
266 |
+
"grad_norm": 0.7250308394432068,
|
267 |
+
"learning_rate": 5e-06,
|
268 |
+
"loss": 0.5648,
|
269 |
+
"step": 360
|
270 |
+
},
|
271 |
+
{
|
272 |
+
"epoch": 1.2802768166089966,
|
273 |
+
"grad_norm": 0.6479423642158508,
|
274 |
+
"learning_rate": 5e-06,
|
275 |
+
"loss": 0.5566,
|
276 |
+
"step": 370
|
277 |
+
},
|
278 |
+
{
|
279 |
+
"epoch": 1.314878892733564,
|
280 |
+
"grad_norm": 0.6422529220581055,
|
281 |
+
"learning_rate": 5e-06,
|
282 |
+
"loss": 0.5596,
|
283 |
+
"step": 380
|
284 |
+
},
|
285 |
+
{
|
286 |
+
"epoch": 1.3494809688581315,
|
287 |
+
"grad_norm": 0.747588038444519,
|
288 |
+
"learning_rate": 5e-06,
|
289 |
+
"loss": 0.5566,
|
290 |
+
"step": 390
|
291 |
+
},
|
292 |
+
{
|
293 |
+
"epoch": 1.3840830449826989,
|
294 |
+
"grad_norm": 0.6396523714065552,
|
295 |
+
"learning_rate": 5e-06,
|
296 |
+
"loss": 0.5582,
|
297 |
+
"step": 400
|
298 |
+
},
|
299 |
+
{
|
300 |
+
"epoch": 1.4186851211072664,
|
301 |
+
"grad_norm": 0.6916091442108154,
|
302 |
+
"learning_rate": 5e-06,
|
303 |
+
"loss": 0.5611,
|
304 |
+
"step": 410
|
305 |
+
},
|
306 |
+
{
|
307 |
+
"epoch": 1.453287197231834,
|
308 |
+
"grad_norm": 0.5232684016227722,
|
309 |
+
"learning_rate": 5e-06,
|
310 |
+
"loss": 0.5552,
|
311 |
+
"step": 420
|
312 |
+
},
|
313 |
+
{
|
314 |
+
"epoch": 1.4878892733564013,
|
315 |
+
"grad_norm": 0.6718825101852417,
|
316 |
+
"learning_rate": 5e-06,
|
317 |
+
"loss": 0.5688,
|
318 |
+
"step": 430
|
319 |
+
},
|
320 |
+
{
|
321 |
+
"epoch": 1.5224913494809689,
|
322 |
+
"grad_norm": 0.5741140842437744,
|
323 |
+
"learning_rate": 5e-06,
|
324 |
+
"loss": 0.5525,
|
325 |
+
"step": 440
|
326 |
+
},
|
327 |
+
{
|
328 |
+
"epoch": 1.5570934256055362,
|
329 |
+
"grad_norm": 0.643785834312439,
|
330 |
+
"learning_rate": 5e-06,
|
331 |
+
"loss": 0.5617,
|
332 |
+
"step": 450
|
333 |
+
},
|
334 |
+
{
|
335 |
+
"epoch": 1.5916955017301038,
|
336 |
+
"grad_norm": 0.7157655954360962,
|
337 |
+
"learning_rate": 5e-06,
|
338 |
+
"loss": 0.5678,
|
339 |
+
"step": 460
|
340 |
+
},
|
341 |
+
{
|
342 |
+
"epoch": 1.6262975778546713,
|
343 |
+
"grad_norm": 0.6625480055809021,
|
344 |
+
"learning_rate": 5e-06,
|
345 |
+
"loss": 0.5608,
|
346 |
+
"step": 470
|
347 |
+
},
|
348 |
+
{
|
349 |
+
"epoch": 1.6608996539792389,
|
350 |
+
"grad_norm": 0.871629536151886,
|
351 |
+
"learning_rate": 5e-06,
|
352 |
+
"loss": 0.5622,
|
353 |
+
"step": 480
|
354 |
+
},
|
355 |
+
{
|
356 |
+
"epoch": 1.6955017301038062,
|
357 |
+
"grad_norm": 0.6766714453697205,
|
358 |
+
"learning_rate": 5e-06,
|
359 |
+
"loss": 0.5498,
|
360 |
+
"step": 490
|
361 |
+
},
|
362 |
+
{
|
363 |
+
"epoch": 1.7301038062283736,
|
364 |
+
"grad_norm": 0.5903865694999695,
|
365 |
+
"learning_rate": 5e-06,
|
366 |
+
"loss": 0.5514,
|
367 |
+
"step": 500
|
368 |
+
},
|
369 |
+
{
|
370 |
+
"epoch": 1.7647058823529411,
|
371 |
+
"grad_norm": 0.7858901023864746,
|
372 |
+
"learning_rate": 5e-06,
|
373 |
+
"loss": 0.5632,
|
374 |
+
"step": 510
|
375 |
+
},
|
376 |
+
{
|
377 |
+
"epoch": 1.7993079584775087,
|
378 |
+
"grad_norm": 0.509254515171051,
|
379 |
+
"learning_rate": 5e-06,
|
380 |
+
"loss": 0.559,
|
381 |
+
"step": 520
|
382 |
+
},
|
383 |
+
{
|
384 |
+
"epoch": 1.8339100346020762,
|
385 |
+
"grad_norm": 0.635212779045105,
|
386 |
+
"learning_rate": 5e-06,
|
387 |
+
"loss": 0.5522,
|
388 |
+
"step": 530
|
389 |
+
},
|
390 |
+
{
|
391 |
+
"epoch": 1.8685121107266436,
|
392 |
+
"grad_norm": 0.5043730735778809,
|
393 |
+
"learning_rate": 5e-06,
|
394 |
+
"loss": 0.5537,
|
395 |
+
"step": 540
|
396 |
+
},
|
397 |
+
{
|
398 |
+
"epoch": 1.903114186851211,
|
399 |
+
"grad_norm": 0.7371939420700073,
|
400 |
+
"learning_rate": 5e-06,
|
401 |
+
"loss": 0.5557,
|
402 |
+
"step": 550
|
403 |
+
},
|
404 |
+
{
|
405 |
+
"epoch": 1.9377162629757785,
|
406 |
+
"grad_norm": 0.5097610354423523,
|
407 |
+
"learning_rate": 5e-06,
|
408 |
+
"loss": 0.5481,
|
409 |
+
"step": 560
|
410 |
+
},
|
411 |
+
{
|
412 |
+
"epoch": 1.972318339100346,
|
413 |
+
"grad_norm": 0.6075087785720825,
|
414 |
+
"learning_rate": 5e-06,
|
415 |
+
"loss": 0.5546,
|
416 |
+
"step": 570
|
417 |
+
},
|
418 |
+
{
|
419 |
+
"epoch": 2.0,
|
420 |
+
"eval_loss": 0.5923461318016052,
|
421 |
+
"eval_runtime": 26.1049,
|
422 |
+
"eval_samples_per_second": 298.22,
|
423 |
+
"eval_steps_per_second": 1.188,
|
424 |
+
"step": 578
|
425 |
+
},
|
426 |
+
{
|
427 |
+
"epoch": 2.0069204152249136,
|
428 |
+
"grad_norm": 1.147298812866211,
|
429 |
+
"learning_rate": 5e-06,
|
430 |
+
"loss": 0.5407,
|
431 |
+
"step": 580
|
432 |
+
},
|
433 |
+
{
|
434 |
+
"epoch": 2.041522491349481,
|
435 |
+
"grad_norm": 0.7817992568016052,
|
436 |
+
"learning_rate": 5e-06,
|
437 |
+
"loss": 0.51,
|
438 |
+
"step": 590
|
439 |
+
},
|
440 |
+
{
|
441 |
+
"epoch": 2.0761245674740483,
|
442 |
+
"grad_norm": 0.6091957092285156,
|
443 |
+
"learning_rate": 5e-06,
|
444 |
+
"loss": 0.5043,
|
445 |
+
"step": 600
|
446 |
+
},
|
447 |
+
{
|
448 |
+
"epoch": 2.110726643598616,
|
449 |
+
"grad_norm": 0.734910786151886,
|
450 |
+
"learning_rate": 5e-06,
|
451 |
+
"loss": 0.508,
|
452 |
+
"step": 610
|
453 |
+
},
|
454 |
+
{
|
455 |
+
"epoch": 2.1453287197231834,
|
456 |
+
"grad_norm": 0.6779142618179321,
|
457 |
+
"learning_rate": 5e-06,
|
458 |
+
"loss": 0.5109,
|
459 |
+
"step": 620
|
460 |
+
},
|
461 |
+
{
|
462 |
+
"epoch": 2.179930795847751,
|
463 |
+
"grad_norm": 0.6702868342399597,
|
464 |
+
"learning_rate": 5e-06,
|
465 |
+
"loss": 0.5151,
|
466 |
+
"step": 630
|
467 |
+
},
|
468 |
+
{
|
469 |
+
"epoch": 2.2145328719723185,
|
470 |
+
"grad_norm": 0.6096834540367126,
|
471 |
+
"learning_rate": 5e-06,
|
472 |
+
"loss": 0.5119,
|
473 |
+
"step": 640
|
474 |
+
},
|
475 |
+
{
|
476 |
+
"epoch": 2.2491349480968856,
|
477 |
+
"grad_norm": 0.7306134700775146,
|
478 |
+
"learning_rate": 5e-06,
|
479 |
+
"loss": 0.5113,
|
480 |
+
"step": 650
|
481 |
+
},
|
482 |
+
{
|
483 |
+
"epoch": 2.283737024221453,
|
484 |
+
"grad_norm": 0.6218770742416382,
|
485 |
+
"learning_rate": 5e-06,
|
486 |
+
"loss": 0.5131,
|
487 |
+
"step": 660
|
488 |
+
},
|
489 |
+
{
|
490 |
+
"epoch": 2.3183391003460208,
|
491 |
+
"grad_norm": 0.5716001987457275,
|
492 |
+
"learning_rate": 5e-06,
|
493 |
+
"loss": 0.5115,
|
494 |
+
"step": 670
|
495 |
+
},
|
496 |
+
{
|
497 |
+
"epoch": 2.3529411764705883,
|
498 |
+
"grad_norm": 0.5372514724731445,
|
499 |
+
"learning_rate": 5e-06,
|
500 |
+
"loss": 0.5188,
|
501 |
+
"step": 680
|
502 |
+
},
|
503 |
+
{
|
504 |
+
"epoch": 2.387543252595156,
|
505 |
+
"grad_norm": 0.5553603172302246,
|
506 |
+
"learning_rate": 5e-06,
|
507 |
+
"loss": 0.5135,
|
508 |
+
"step": 690
|
509 |
+
},
|
510 |
+
{
|
511 |
+
"epoch": 2.422145328719723,
|
512 |
+
"grad_norm": 0.6243720054626465,
|
513 |
+
"learning_rate": 5e-06,
|
514 |
+
"loss": 0.5182,
|
515 |
+
"step": 700
|
516 |
+
},
|
517 |
+
{
|
518 |
+
"epoch": 2.4567474048442905,
|
519 |
+
"grad_norm": 0.6500486135482788,
|
520 |
+
"learning_rate": 5e-06,
|
521 |
+
"loss": 0.5085,
|
522 |
+
"step": 710
|
523 |
+
},
|
524 |
+
{
|
525 |
+
"epoch": 2.491349480968858,
|
526 |
+
"grad_norm": 0.6719836592674255,
|
527 |
+
"learning_rate": 5e-06,
|
528 |
+
"loss": 0.514,
|
529 |
+
"step": 720
|
530 |
+
},
|
531 |
+
{
|
532 |
+
"epoch": 2.5259515570934257,
|
533 |
+
"grad_norm": 0.6871230602264404,
|
534 |
+
"learning_rate": 5e-06,
|
535 |
+
"loss": 0.5094,
|
536 |
+
"step": 730
|
537 |
+
},
|
538 |
+
{
|
539 |
+
"epoch": 2.5605536332179932,
|
540 |
+
"grad_norm": 0.8423399925231934,
|
541 |
+
"learning_rate": 5e-06,
|
542 |
+
"loss": 0.5138,
|
543 |
+
"step": 740
|
544 |
+
},
|
545 |
+
{
|
546 |
+
"epoch": 2.595155709342561,
|
547 |
+
"grad_norm": 0.557617723941803,
|
548 |
+
"learning_rate": 5e-06,
|
549 |
+
"loss": 0.5192,
|
550 |
+
"step": 750
|
551 |
+
},
|
552 |
+
{
|
553 |
+
"epoch": 2.629757785467128,
|
554 |
+
"grad_norm": 0.7015785574913025,
|
555 |
+
"learning_rate": 5e-06,
|
556 |
+
"loss": 0.5122,
|
557 |
+
"step": 760
|
558 |
+
},
|
559 |
+
{
|
560 |
+
"epoch": 2.6643598615916955,
|
561 |
+
"grad_norm": 0.6643463373184204,
|
562 |
+
"learning_rate": 5e-06,
|
563 |
+
"loss": 0.5136,
|
564 |
+
"step": 770
|
565 |
+
},
|
566 |
+
{
|
567 |
+
"epoch": 2.698961937716263,
|
568 |
+
"grad_norm": 0.8234036564826965,
|
569 |
+
"learning_rate": 5e-06,
|
570 |
+
"loss": 0.5147,
|
571 |
+
"step": 780
|
572 |
+
},
|
573 |
+
{
|
574 |
+
"epoch": 2.7335640138408306,
|
575 |
+
"grad_norm": 0.5616887211799622,
|
576 |
+
"learning_rate": 5e-06,
|
577 |
+
"loss": 0.5147,
|
578 |
+
"step": 790
|
579 |
+
},
|
580 |
+
{
|
581 |
+
"epoch": 2.7681660899653977,
|
582 |
+
"grad_norm": 0.6399916410446167,
|
583 |
+
"learning_rate": 5e-06,
|
584 |
+
"loss": 0.5104,
|
585 |
+
"step": 800
|
586 |
+
},
|
587 |
+
{
|
588 |
+
"epoch": 2.8027681660899653,
|
589 |
+
"grad_norm": 0.5685309767723083,
|
590 |
+
"learning_rate": 5e-06,
|
591 |
+
"loss": 0.5128,
|
592 |
+
"step": 810
|
593 |
+
},
|
594 |
+
{
|
595 |
+
"epoch": 2.837370242214533,
|
596 |
+
"grad_norm": 0.5425155758857727,
|
597 |
+
"learning_rate": 5e-06,
|
598 |
+
"loss": 0.5132,
|
599 |
+
"step": 820
|
600 |
+
},
|
601 |
+
{
|
602 |
+
"epoch": 2.8719723183391004,
|
603 |
+
"grad_norm": 0.8712612986564636,
|
604 |
+
"learning_rate": 5e-06,
|
605 |
+
"loss": 0.5074,
|
606 |
+
"step": 830
|
607 |
+
},
|
608 |
+
{
|
609 |
+
"epoch": 2.906574394463668,
|
610 |
+
"grad_norm": 0.8342138528823853,
|
611 |
+
"learning_rate": 5e-06,
|
612 |
+
"loss": 0.5215,
|
613 |
+
"step": 840
|
614 |
+
},
|
615 |
+
{
|
616 |
+
"epoch": 2.9411764705882355,
|
617 |
+
"grad_norm": 0.7109184861183167,
|
618 |
+
"learning_rate": 5e-06,
|
619 |
+
"loss": 0.5102,
|
620 |
+
"step": 850
|
621 |
+
},
|
622 |
+
{
|
623 |
+
"epoch": 2.9757785467128026,
|
624 |
+
"grad_norm": 0.6473217606544495,
|
625 |
+
"learning_rate": 5e-06,
|
626 |
+
"loss": 0.5168,
|
627 |
+
"step": 860
|
628 |
+
},
|
629 |
+
{
|
630 |
+
"epoch": 3.0,
|
631 |
+
"eval_loss": 0.5936245918273926,
|
632 |
+
"eval_runtime": 26.2859,
|
633 |
+
"eval_samples_per_second": 296.166,
|
634 |
+
"eval_steps_per_second": 1.179,
|
635 |
+
"step": 867
|
636 |
+
},
|
637 |
+
{
|
638 |
+
"epoch": 3.0,
|
639 |
+
"step": 867,
|
640 |
+
"total_flos": 4.093705665380352e+19,
|
641 |
+
"train_loss": 0.5755822655788893,
|
642 |
+
"train_runtime": 6372.9887,
|
643 |
+
"train_samples_per_second": 69.622,
|
644 |
+
"train_steps_per_second": 0.136
|
645 |
+
}
|
646 |
+
],
|
647 |
+
"logging_steps": 10,
|
648 |
+
"max_steps": 867,
|
649 |
+
"num_input_tokens_seen": 0,
|
650 |
+
"num_train_epochs": 3,
|
651 |
+
"save_steps": 500,
|
652 |
+
"stateful_callbacks": {
|
653 |
+
"TrainerControl": {
|
654 |
+
"args": {
|
655 |
+
"should_epoch_stop": false,
|
656 |
+
"should_evaluate": false,
|
657 |
+
"should_log": false,
|
658 |
+
"should_save": true,
|
659 |
+
"should_training_stop": true
|
660 |
+
},
|
661 |
+
"attributes": {}
|
662 |
+
}
|
663 |
+
},
|
664 |
+
"total_flos": 4.093705665380352e+19,
|
665 |
+
"train_batch_size": 16,
|
666 |
+
"trial_name": null,
|
667 |
+
"trial_params": null
|
668 |
+
}
|
training_eval_loss.png
ADDED
training_loss.png
ADDED