Model save
Browse files- adapter_config.json +2 -2
- adapter_model.safetensors +1 -1
- all_results.json +7 -7
- eval_results.json +4 -4
- train_results.json +3 -3
- trainer_state.json +10 -10
- training_args.bin +1 -1
adapter_config.json
CHANGED
@@ -17,9 +17,9 @@
|
|
17 |
"revision": null,
|
18 |
"target_modules": [
|
19 |
"k_proj",
|
20 |
-
"v_proj",
|
21 |
"q_proj",
|
22 |
-
"o_proj"
|
|
|
23 |
],
|
24 |
"task_type": "CAUSAL_LM"
|
25 |
}
|
|
|
17 |
"revision": null,
|
18 |
"target_modules": [
|
19 |
"k_proj",
|
|
|
20 |
"q_proj",
|
21 |
+
"o_proj",
|
22 |
+
"v_proj"
|
23 |
],
|
24 |
"task_type": "CAUSAL_LM"
|
25 |
}
|
adapter_model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 109086672
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:48b34a428074e5b7cc32b3fd5343936b35b0a5c869b9ecd3c09f7000dcb50b11
|
3 |
size 109086672
|
all_results.json
CHANGED
@@ -1,13 +1,13 @@
|
|
1 |
{
|
2 |
"epoch": 0.67,
|
3 |
-
"eval_loss": 1.
|
4 |
-
"eval_runtime":
|
5 |
"eval_samples": 23110,
|
6 |
-
"eval_samples_per_second": 9.
|
7 |
-
"eval_steps_per_second": 0.
|
8 |
-
"train_loss": 1.
|
9 |
-
"train_runtime":
|
10 |
"train_samples": 207865,
|
11 |
-
"train_samples_per_second": 2.
|
12 |
"train_steps_per_second": 0.006
|
13 |
}
|
|
|
1 |
{
|
2 |
"epoch": 0.67,
|
3 |
+
"eval_loss": 1.0666700601577759,
|
4 |
+
"eval_runtime": 2384.3246,
|
5 |
"eval_samples": 23110,
|
6 |
+
"eval_samples_per_second": 9.692,
|
7 |
+
"eval_steps_per_second": 0.152,
|
8 |
+
"train_loss": 1.1599246359923308,
|
9 |
+
"train_runtime": 72626.7254,
|
10 |
"train_samples": 207865,
|
11 |
+
"train_samples_per_second": 2.862,
|
12 |
"train_steps_per_second": 0.006
|
13 |
}
|
eval_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
"epoch": 0.67,
|
3 |
-
"eval_loss": 1.
|
4 |
-
"eval_runtime":
|
5 |
"eval_samples": 23110,
|
6 |
-
"eval_samples_per_second": 9.
|
7 |
-
"eval_steps_per_second": 0.
|
8 |
}
|
|
|
1 |
{
|
2 |
"epoch": 0.67,
|
3 |
+
"eval_loss": 1.0666700601577759,
|
4 |
+
"eval_runtime": 2384.3246,
|
5 |
"eval_samples": 23110,
|
6 |
+
"eval_samples_per_second": 9.692,
|
7 |
+
"eval_steps_per_second": 0.152
|
8 |
}
|
train_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
"epoch": 0.67,
|
3 |
-
"train_loss": 1.
|
4 |
-
"train_runtime":
|
5 |
"train_samples": 207865,
|
6 |
-
"train_samples_per_second": 2.
|
7 |
"train_steps_per_second": 0.006
|
8 |
}
|
|
|
1 |
{
|
2 |
"epoch": 0.67,
|
3 |
+
"train_loss": 1.1599246359923308,
|
4 |
+
"train_runtime": 72626.7254,
|
5 |
"train_samples": 207865,
|
6 |
+
"train_samples_per_second": 2.862,
|
7 |
"train_steps_per_second": 0.006
|
8 |
}
|
trainer_state.json
CHANGED
@@ -77,7 +77,7 @@
|
|
77 |
{
|
78 |
"epoch": 0.14,
|
79 |
"learning_rate": 1.910797282022027e-05,
|
80 |
-
"loss": 1.
|
81 |
"step": 55
|
82 |
},
|
83 |
{
|
@@ -149,7 +149,7 @@
|
|
149 |
{
|
150 |
"epoch": 0.28,
|
151 |
"learning_rate": 1.629520819706912e-05,
|
152 |
-
"loss": 1.
|
153 |
"step": 115
|
154 |
},
|
155 |
{
|
@@ -179,7 +179,7 @@
|
|
179 |
{
|
180 |
"epoch": 0.34,
|
181 |
"learning_rate": 1.4684084406997903e-05,
|
182 |
-
"loss": 1.
|
183 |
"step": 140
|
184 |
},
|
185 |
{
|
@@ -275,7 +275,7 @@
|
|
275 |
{
|
276 |
"epoch": 0.54,
|
277 |
"learning_rate": 8.688345254588579e-06,
|
278 |
-
"loss": 1.
|
279 |
"step": 220
|
280 |
},
|
281 |
{
|
@@ -340,9 +340,9 @@
|
|
340 |
},
|
341 |
{
|
342 |
"epoch": 0.67,
|
343 |
-
"eval_loss": 1.
|
344 |
-
"eval_runtime":
|
345 |
-
"eval_samples_per_second": 9.
|
346 |
"eval_steps_per_second": 0.152,
|
347 |
"step": 272
|
348 |
},
|
@@ -350,9 +350,9 @@
|
|
350 |
"epoch": 0.67,
|
351 |
"step": 272,
|
352 |
"total_flos": 4.932776606721638e+16,
|
353 |
-
"train_loss": 1.
|
354 |
-
"train_runtime":
|
355 |
-
"train_samples_per_second": 2.
|
356 |
"train_steps_per_second": 0.006
|
357 |
}
|
358 |
],
|
|
|
77 |
{
|
78 |
"epoch": 0.14,
|
79 |
"learning_rate": 1.910797282022027e-05,
|
80 |
+
"loss": 1.2334,
|
81 |
"step": 55
|
82 |
},
|
83 |
{
|
|
|
149 |
{
|
150 |
"epoch": 0.28,
|
151 |
"learning_rate": 1.629520819706912e-05,
|
152 |
+
"loss": 1.107,
|
153 |
"step": 115
|
154 |
},
|
155 |
{
|
|
|
179 |
{
|
180 |
"epoch": 0.34,
|
181 |
"learning_rate": 1.4684084406997903e-05,
|
182 |
+
"loss": 1.091,
|
183 |
"step": 140
|
184 |
},
|
185 |
{
|
|
|
275 |
{
|
276 |
"epoch": 0.54,
|
277 |
"learning_rate": 8.688345254588579e-06,
|
278 |
+
"loss": 1.0686,
|
279 |
"step": 220
|
280 |
},
|
281 |
{
|
|
|
340 |
},
|
341 |
{
|
342 |
"epoch": 0.67,
|
343 |
+
"eval_loss": 1.066724419593811,
|
344 |
+
"eval_runtime": 2385.329,
|
345 |
+
"eval_samples_per_second": 9.688,
|
346 |
"eval_steps_per_second": 0.152,
|
347 |
"step": 272
|
348 |
},
|
|
|
350 |
"epoch": 0.67,
|
351 |
"step": 272,
|
352 |
"total_flos": 4.932776606721638e+16,
|
353 |
+
"train_loss": 1.1599246359923308,
|
354 |
+
"train_runtime": 72626.7254,
|
355 |
+
"train_samples_per_second": 2.862,
|
356 |
"train_steps_per_second": 0.006
|
357 |
}
|
358 |
],
|
training_args.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 5624
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9a116181e8e43014f8554b0cc2a81c8131193600bb4d374a77cd396fcfdc6416
|
3 |
size 5624
|