Training in progress, step 3, checkpoint
Browse files
last-checkpoint/adapter_config.json
CHANGED
@@ -20,13 +20,13 @@
|
|
20 |
"rank_pattern": {},
|
21 |
"revision": null,
|
22 |
"target_modules": [
|
|
|
|
|
|
|
23 |
"q_proj",
|
24 |
-
"up_proj",
|
25 |
"k_proj",
|
26 |
-
"gate_proj",
|
27 |
"v_proj",
|
28 |
-
"
|
29 |
-
"o_proj"
|
30 |
],
|
31 |
"task_type": "CAUSAL_LM",
|
32 |
"use_dora": false,
|
|
|
20 |
"rank_pattern": {},
|
21 |
"revision": null,
|
22 |
"target_modules": [
|
23 |
+
"down_proj",
|
24 |
+
"o_proj",
|
25 |
+
"gate_proj",
|
26 |
"q_proj",
|
|
|
27 |
"k_proj",
|
|
|
28 |
"v_proj",
|
29 |
+
"up_proj"
|
|
|
30 |
],
|
31 |
"task_type": "CAUSAL_LM",
|
32 |
"use_dora": false,
|
last-checkpoint/adapter_model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 35237104
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fd524b06507682b293f3f3df66c98ad864abf4ee62dfc80b4471e21d05147f69
|
3 |
size 35237104
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 18810036
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:051d7e8d7885c7935b84eb950597652be5f384e97da94946990c7c750627354f
|
3 |
size 18810036
|
last-checkpoint/trainer_state.json
CHANGED
@@ -10,7 +10,7 @@
|
|
10 |
"log_history": [
|
11 |
{
|
12 |
"epoch": 0.0005163955589981926,
|
13 |
-
"grad_norm": 1.
|
14 |
"learning_rate": 2e-05,
|
15 |
"loss": 2.0729,
|
16 |
"step": 1
|
@@ -18,31 +18,31 @@
|
|
18 |
{
|
19 |
"epoch": 0.0005163955589981926,
|
20 |
"eval_loss": 2.127951145172119,
|
21 |
-
"eval_runtime":
|
22 |
-
"eval_samples_per_second": 1.
|
23 |
-
"eval_steps_per_second": 0.
|
24 |
"step": 1
|
25 |
},
|
26 |
{
|
27 |
"epoch": 0.0010327911179963851,
|
28 |
-
"grad_norm": 0.
|
29 |
"learning_rate": 4e-05,
|
30 |
"loss": 1.9541,
|
31 |
"step": 2
|
32 |
},
|
33 |
{
|
34 |
"epoch": 0.001549186676994578,
|
35 |
-
"grad_norm": 0.
|
36 |
"learning_rate": 6e-05,
|
37 |
-
"loss": 2.
|
38 |
"step": 3
|
39 |
},
|
40 |
{
|
41 |
"epoch": 0.001549186676994578,
|
42 |
-
"eval_loss": 2.
|
43 |
-
"eval_runtime":
|
44 |
-
"eval_samples_per_second": 1.
|
45 |
-
"eval_steps_per_second": 0.
|
46 |
"step": 3
|
47 |
}
|
48 |
],
|
|
|
10 |
"log_history": [
|
11 |
{
|
12 |
"epoch": 0.0005163955589981926,
|
13 |
+
"grad_norm": 1.1844890117645264,
|
14 |
"learning_rate": 2e-05,
|
15 |
"loss": 2.0729,
|
16 |
"step": 1
|
|
|
18 |
{
|
19 |
"epoch": 0.0005163955589981926,
|
20 |
"eval_loss": 2.127951145172119,
|
21 |
+
"eval_runtime": 542.3791,
|
22 |
+
"eval_samples_per_second": 1.504,
|
23 |
+
"eval_steps_per_second": 0.752,
|
24 |
"step": 1
|
25 |
},
|
26 |
{
|
27 |
"epoch": 0.0010327911179963851,
|
28 |
+
"grad_norm": 0.7695030570030212,
|
29 |
"learning_rate": 4e-05,
|
30 |
"loss": 1.9541,
|
31 |
"step": 2
|
32 |
},
|
33 |
{
|
34 |
"epoch": 0.001549186676994578,
|
35 |
+
"grad_norm": 0.9567698836326599,
|
36 |
"learning_rate": 6e-05,
|
37 |
+
"loss": 2.0493,
|
38 |
"step": 3
|
39 |
},
|
40 |
{
|
41 |
"epoch": 0.001549186676994578,
|
42 |
+
"eval_loss": 2.122117519378662,
|
43 |
+
"eval_runtime": 542.3452,
|
44 |
+
"eval_samples_per_second": 1.505,
|
45 |
+
"eval_steps_per_second": 0.752,
|
46 |
"step": 3
|
47 |
}
|
48 |
],
|
last-checkpoint/training_args.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 6776
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:32216e2a624352f583504f70daf3f888fc0cd4905fd9cee1127453e25379faf8
|
3 |
size 6776
|