Upload 3 files
Browse filesadd trained lora
- adapter_config.json +17 -0
- adapter_model.bin +3 -0
- trainer_state.json +341 -0
adapter_config.json
ADDED
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"base_model_name_or_path": "/home/llmadmin/models/my_llama_hf/llama_hf_7B",
|
3 |
+
"bias": "none",
|
4 |
+
"fan_in_fan_out": false,
|
5 |
+
"inference_mode": true,
|
6 |
+
"init_lora_weights": true,
|
7 |
+
"lora_alpha": 64,
|
8 |
+
"lora_dropout": 0.05,
|
9 |
+
"modules_to_save": null,
|
10 |
+
"peft_type": "LORA",
|
11 |
+
"r": 32,
|
12 |
+
"target_modules": [
|
13 |
+
"q_proj",
|
14 |
+
"v_proj"
|
15 |
+
],
|
16 |
+
"task_type": "CAUSAL_LM"
|
17 |
+
}
|
adapter_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9a74aee00df877f98d1fa126b04fd13f89db5f962a1c4b80d95a9f7ffea87bf9
|
3 |
+
size 67154893
|
trainer_state.json
ADDED
@@ -0,0 +1,341 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 1.8454276323318481,
|
3 |
+
"best_model_checkpoint": "/home/llmadmin/lawrence/autoWSB/data/tunned_v4/checkpoint-1000",
|
4 |
+
"epoch": 2.9994666666666667,
|
5 |
+
"global_step": 4218,
|
6 |
+
"is_hyper_param_search": false,
|
7 |
+
"is_local_process_zero": true,
|
8 |
+
"is_world_process_zero": true,
|
9 |
+
"log_history": [
|
10 |
+
{
|
11 |
+
"epoch": 0.07,
|
12 |
+
"learning_rate": 8.333333333333333e-05,
|
13 |
+
"loss": 1.9887,
|
14 |
+
"step": 100
|
15 |
+
},
|
16 |
+
{
|
17 |
+
"epoch": 0.14,
|
18 |
+
"learning_rate": 0.00016666666666666666,
|
19 |
+
"loss": 1.7819,
|
20 |
+
"step": 200
|
21 |
+
},
|
22 |
+
{
|
23 |
+
"epoch": 0.21,
|
24 |
+
"learning_rate": 0.00025,
|
25 |
+
"loss": 1.7572,
|
26 |
+
"step": 300
|
27 |
+
},
|
28 |
+
{
|
29 |
+
"epoch": 0.28,
|
30 |
+
"learning_rate": 0.00024400444104134763,
|
31 |
+
"loss": 3.1654,
|
32 |
+
"step": 400
|
33 |
+
},
|
34 |
+
{
|
35 |
+
"epoch": 0.36,
|
36 |
+
"learning_rate": 0.00023762618683001533,
|
37 |
+
"loss": 2.402,
|
38 |
+
"step": 500
|
39 |
+
},
|
40 |
+
{
|
41 |
+
"epoch": 0.36,
|
42 |
+
"eval_loss": 2.0042166709899902,
|
43 |
+
"eval_runtime": 1078.6167,
|
44 |
+
"eval_samples_per_second": 23.178,
|
45 |
+
"eval_steps_per_second": 0.579,
|
46 |
+
"step": 500
|
47 |
+
},
|
48 |
+
{
|
49 |
+
"epoch": 0.43,
|
50 |
+
"learning_rate": 0.000231247932618683,
|
51 |
+
"loss": 1.9543,
|
52 |
+
"step": 600
|
53 |
+
},
|
54 |
+
{
|
55 |
+
"epoch": 0.5,
|
56 |
+
"learning_rate": 0.0002248696784073507,
|
57 |
+
"loss": 1.905,
|
58 |
+
"step": 700
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"epoch": 0.57,
|
62 |
+
"learning_rate": 0.00021849142419601836,
|
63 |
+
"loss": 1.8835,
|
64 |
+
"step": 800
|
65 |
+
},
|
66 |
+
{
|
67 |
+
"epoch": 0.64,
|
68 |
+
"learning_rate": 0.00021211316998468607,
|
69 |
+
"loss": 1.8628,
|
70 |
+
"step": 900
|
71 |
+
},
|
72 |
+
{
|
73 |
+
"epoch": 0.71,
|
74 |
+
"learning_rate": 0.00020573491577335375,
|
75 |
+
"loss": 1.8463,
|
76 |
+
"step": 1000
|
77 |
+
},
|
78 |
+
{
|
79 |
+
"epoch": 0.71,
|
80 |
+
"eval_loss": 1.8454276323318481,
|
81 |
+
"eval_runtime": 1080.8608,
|
82 |
+
"eval_samples_per_second": 23.13,
|
83 |
+
"eval_steps_per_second": 0.578,
|
84 |
+
"step": 1000
|
85 |
+
},
|
86 |
+
{
|
87 |
+
"epoch": 0.78,
|
88 |
+
"learning_rate": 0.00019935666156202145,
|
89 |
+
"loss": 1.8654,
|
90 |
+
"step": 1100
|
91 |
+
},
|
92 |
+
{
|
93 |
+
"epoch": 0.85,
|
94 |
+
"learning_rate": 0.00019297840735068913,
|
95 |
+
"loss": 1.8511,
|
96 |
+
"step": 1200
|
97 |
+
},
|
98 |
+
{
|
99 |
+
"epoch": 0.92,
|
100 |
+
"learning_rate": 0.00018660015313935684,
|
101 |
+
"loss": 1.8424,
|
102 |
+
"step": 1300
|
103 |
+
},
|
104 |
+
{
|
105 |
+
"epoch": 1.0,
|
106 |
+
"learning_rate": 0.0001802218989280245,
|
107 |
+
"loss": 1.8579,
|
108 |
+
"step": 1400
|
109 |
+
},
|
110 |
+
{
|
111 |
+
"epoch": 1.07,
|
112 |
+
"learning_rate": 0.00017843598774885147,
|
113 |
+
"loss": 23.4765,
|
114 |
+
"step": 1500
|
115 |
+
},
|
116 |
+
{
|
117 |
+
"epoch": 1.07,
|
118 |
+
"eval_loss": 10.320993423461914,
|
119 |
+
"eval_runtime": 1165.3618,
|
120 |
+
"eval_samples_per_second": 21.453,
|
121 |
+
"eval_steps_per_second": 0.536,
|
122 |
+
"step": 1500
|
123 |
+
},
|
124 |
+
{
|
125 |
+
"epoch": 1.14,
|
126 |
+
"learning_rate": 0.00017767059724349159,
|
127 |
+
"loss": 26.8324,
|
128 |
+
"step": 1600
|
129 |
+
},
|
130 |
+
{
|
131 |
+
"epoch": 1.21,
|
132 |
+
"learning_rate": 0.00017129234303215926,
|
133 |
+
"loss": 33.5803,
|
134 |
+
"step": 1700
|
135 |
+
},
|
136 |
+
{
|
137 |
+
"epoch": 1.28,
|
138 |
+
"learning_rate": 0.00016491408882082697,
|
139 |
+
"loss": 75831.04,
|
140 |
+
"step": 1800
|
141 |
+
},
|
142 |
+
{
|
143 |
+
"epoch": 1.35,
|
144 |
+
"learning_rate": 0.00015853583460949465,
|
145 |
+
"loss": 112434.02,
|
146 |
+
"step": 1900
|
147 |
+
},
|
148 |
+
{
|
149 |
+
"epoch": 1.42,
|
150 |
+
"learning_rate": 0.00015215758039816232,
|
151 |
+
"loss": 9892.82,
|
152 |
+
"step": 2000
|
153 |
+
},
|
154 |
+
{
|
155 |
+
"epoch": 1.42,
|
156 |
+
"eval_loss": 10.320993423461914,
|
157 |
+
"eval_runtime": 1166.8445,
|
158 |
+
"eval_samples_per_second": 21.425,
|
159 |
+
"eval_steps_per_second": 0.536,
|
160 |
+
"step": 2000
|
161 |
+
},
|
162 |
+
{
|
163 |
+
"epoch": 1.49,
|
164 |
+
"learning_rate": 0.00014577932618683,
|
165 |
+
"loss": 9040.6219,
|
166 |
+
"step": 2100
|
167 |
+
},
|
168 |
+
{
|
169 |
+
"epoch": 1.56,
|
170 |
+
"learning_rate": 0.0001394010719754977,
|
171 |
+
"loss": 2015.4547,
|
172 |
+
"step": 2200
|
173 |
+
},
|
174 |
+
{
|
175 |
+
"epoch": 1.64,
|
176 |
+
"learning_rate": 0.0001330228177641654,
|
177 |
+
"loss": 16854.6437,
|
178 |
+
"step": 2300
|
179 |
+
},
|
180 |
+
{
|
181 |
+
"epoch": 1.71,
|
182 |
+
"learning_rate": 0.0001266445635528331,
|
183 |
+
"loss": 795.047,
|
184 |
+
"step": 2400
|
185 |
+
},
|
186 |
+
{
|
187 |
+
"epoch": 1.78,
|
188 |
+
"learning_rate": 0.00012026630934150074,
|
189 |
+
"loss": 38857.9575,
|
190 |
+
"step": 2500
|
191 |
+
},
|
192 |
+
{
|
193 |
+
"epoch": 1.78,
|
194 |
+
"eval_loss": 10.320993423461914,
|
195 |
+
"eval_runtime": 1167.9826,
|
196 |
+
"eval_samples_per_second": 21.404,
|
197 |
+
"eval_steps_per_second": 0.535,
|
198 |
+
"step": 2500
|
199 |
+
},
|
200 |
+
{
|
201 |
+
"epoch": 1.85,
|
202 |
+
"learning_rate": 0.00011388805513016846,
|
203 |
+
"loss": 185.0979,
|
204 |
+
"step": 2600
|
205 |
+
},
|
206 |
+
{
|
207 |
+
"epoch": 1.92,
|
208 |
+
"learning_rate": 0.00010750980091883614,
|
209 |
+
"loss": 12796.2138,
|
210 |
+
"step": 2700
|
211 |
+
},
|
212 |
+
{
|
213 |
+
"epoch": 1.99,
|
214 |
+
"learning_rate": 0.00010113154670750384,
|
215 |
+
"loss": 14625.28,
|
216 |
+
"step": 2800
|
217 |
+
},
|
218 |
+
{
|
219 |
+
"epoch": 2.06,
|
220 |
+
"learning_rate": 9.475329249617151e-05,
|
221 |
+
"loss": 271.5516,
|
222 |
+
"step": 2900
|
223 |
+
},
|
224 |
+
{
|
225 |
+
"epoch": 2.13,
|
226 |
+
"learning_rate": 8.837503828483921e-05,
|
227 |
+
"loss": 5358.3219,
|
228 |
+
"step": 3000
|
229 |
+
},
|
230 |
+
{
|
231 |
+
"epoch": 2.13,
|
232 |
+
"eval_loss": 10.320993423461914,
|
233 |
+
"eval_runtime": 1167.2296,
|
234 |
+
"eval_samples_per_second": 21.418,
|
235 |
+
"eval_steps_per_second": 0.535,
|
236 |
+
"step": 3000
|
237 |
+
},
|
238 |
+
{
|
239 |
+
"epoch": 2.2,
|
240 |
+
"learning_rate": 8.199678407350689e-05,
|
241 |
+
"loss": 106711.13,
|
242 |
+
"step": 3100
|
243 |
+
},
|
244 |
+
{
|
245 |
+
"epoch": 2.28,
|
246 |
+
"learning_rate": 7.561852986217458e-05,
|
247 |
+
"loss": 141536.53,
|
248 |
+
"step": 3200
|
249 |
+
},
|
250 |
+
{
|
251 |
+
"epoch": 2.35,
|
252 |
+
"learning_rate": 6.924027565084226e-05,
|
253 |
+
"loss": 784.2875,
|
254 |
+
"step": 3300
|
255 |
+
},
|
256 |
+
{
|
257 |
+
"epoch": 2.42,
|
258 |
+
"learning_rate": 6.286202143950997e-05,
|
259 |
+
"loss": 95102.5,
|
260 |
+
"step": 3400
|
261 |
+
},
|
262 |
+
{
|
263 |
+
"epoch": 2.49,
|
264 |
+
"learning_rate": 5.648376722817764e-05,
|
265 |
+
"loss": 2327.5258,
|
266 |
+
"step": 3500
|
267 |
+
},
|
268 |
+
{
|
269 |
+
"epoch": 2.49,
|
270 |
+
"eval_loss": 10.320993423461914,
|
271 |
+
"eval_runtime": 1166.8332,
|
272 |
+
"eval_samples_per_second": 21.426,
|
273 |
+
"eval_steps_per_second": 0.536,
|
274 |
+
"step": 3500
|
275 |
+
},
|
276 |
+
{
|
277 |
+
"epoch": 2.56,
|
278 |
+
"learning_rate": 5.010551301684534e-05,
|
279 |
+
"loss": 59017.795,
|
280 |
+
"step": 3600
|
281 |
+
},
|
282 |
+
{
|
283 |
+
"epoch": 2.63,
|
284 |
+
"learning_rate": 4.3727258805513014e-05,
|
285 |
+
"loss": 680382.64,
|
286 |
+
"step": 3700
|
287 |
+
},
|
288 |
+
{
|
289 |
+
"epoch": 2.7,
|
290 |
+
"learning_rate": 3.734900459418072e-05,
|
291 |
+
"loss": 1483.1705,
|
292 |
+
"step": 3800
|
293 |
+
},
|
294 |
+
{
|
295 |
+
"epoch": 2.77,
|
296 |
+
"learning_rate": 3.097075038284839e-05,
|
297 |
+
"loss": 30906.0775,
|
298 |
+
"step": 3900
|
299 |
+
},
|
300 |
+
{
|
301 |
+
"epoch": 2.84,
|
302 |
+
"learning_rate": 2.459249617151609e-05,
|
303 |
+
"loss": 46132.635,
|
304 |
+
"step": 4000
|
305 |
+
},
|
306 |
+
{
|
307 |
+
"epoch": 2.84,
|
308 |
+
"eval_loss": 10.320993423461914,
|
309 |
+
"eval_runtime": 1166.5367,
|
310 |
+
"eval_samples_per_second": 21.431,
|
311 |
+
"eval_steps_per_second": 0.536,
|
312 |
+
"step": 4000
|
313 |
+
},
|
314 |
+
{
|
315 |
+
"epoch": 2.92,
|
316 |
+
"learning_rate": 1.8214241960183767e-05,
|
317 |
+
"loss": 31202.95,
|
318 |
+
"step": 4100
|
319 |
+
},
|
320 |
+
{
|
321 |
+
"epoch": 2.99,
|
322 |
+
"learning_rate": 1.1835987748851469e-05,
|
323 |
+
"loss": 2477.5823,
|
324 |
+
"step": 4200
|
325 |
+
},
|
326 |
+
{
|
327 |
+
"epoch": 3.0,
|
328 |
+
"step": 4218,
|
329 |
+
"total_flos": 2.7466670484779696e+19,
|
330 |
+
"train_loss": 35493.948473290504,
|
331 |
+
"train_runtime": 121039.6951,
|
332 |
+
"train_samples_per_second": 5.577,
|
333 |
+
"train_steps_per_second": 0.035
|
334 |
+
}
|
335 |
+
],
|
336 |
+
"max_steps": 4218,
|
337 |
+
"num_train_epochs": 3,
|
338 |
+
"total_flos": 2.7466670484779696e+19,
|
339 |
+
"trial_name": null,
|
340 |
+
"trial_params": null
|
341 |
+
}
|