Training in progress, step 1250, checkpoint
Browse files
checkpoint-1250/adapter_config.json
CHANGED
@@ -27,11 +27,11 @@
|
|
27 |
"revision": null,
|
28 |
"target_modules": [
|
29 |
"k_proj",
|
30 |
-
"down_proj",
|
31 |
-
"v_proj",
|
32 |
-
"o_proj",
|
33 |
"gate_proj",
|
34 |
"q_proj",
|
|
|
|
|
|
|
35 |
"up_proj"
|
36 |
],
|
37 |
"task_type": null,
|
|
|
27 |
"revision": null,
|
28 |
"target_modules": [
|
29 |
"k_proj",
|
|
|
|
|
|
|
30 |
"gate_proj",
|
31 |
"q_proj",
|
32 |
+
"v_proj",
|
33 |
+
"o_proj",
|
34 |
+
"down_proj",
|
35 |
"up_proj"
|
36 |
],
|
37 |
"task_type": null,
|
checkpoint-1250/adapter_model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 400616360
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f3ed338e03e0c16ea7eb67ec41f71a962a2ba38942ff0b78cf603a6cd97172e6
|
3 |
size 400616360
|
checkpoint-1250/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 205100562
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:787ca7f42cb3fec702d3291bd5904451718c598f09a17b9f3f50027ca9ae7814
|
3 |
size 205100562
|
checkpoint-1250/rng_state.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 14308
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b6e269ce06d53a33198330136e8c427552a61f30abb8ee02be6c46f12f8f6467
|
3 |
size 14308
|
checkpoint-1250/trainer_state.json
CHANGED
@@ -70,17 +70,17 @@
|
|
70 |
},
|
71 |
{
|
72 |
"epoch": 1.167114015637764,
|
73 |
-
"grad_norm": 2.
|
74 |
"learning_rate": 3.915853581228413e-05,
|
75 |
-
"loss": 0.
|
76 |
"step": 1250
|
77 |
},
|
78 |
{
|
79 |
"epoch": 1.167114015637764,
|
80 |
-
"eval_loss": 0.
|
81 |
-
"eval_runtime":
|
82 |
-
"eval_samples_per_second": 3.
|
83 |
-
"eval_steps_per_second": 3.
|
84 |
"step": 1250
|
85 |
}
|
86 |
],
|
@@ -101,7 +101,7 @@
|
|
101 |
"attributes": {}
|
102 |
}
|
103 |
},
|
104 |
-
"total_flos": 1.
|
105 |
"train_batch_size": 1,
|
106 |
"trial_name": null,
|
107 |
"trial_params": null
|
|
|
70 |
},
|
71 |
{
|
72 |
"epoch": 1.167114015637764,
|
73 |
+
"grad_norm": 2.2942659854888916,
|
74 |
"learning_rate": 3.915853581228413e-05,
|
75 |
+
"loss": 0.7993,
|
76 |
"step": 1250
|
77 |
},
|
78 |
{
|
79 |
"epoch": 1.167114015637764,
|
80 |
+
"eval_loss": 0.1208883598446846,
|
81 |
+
"eval_runtime": 21.9734,
|
82 |
+
"eval_samples_per_second": 3.959,
|
83 |
+
"eval_steps_per_second": 3.959,
|
84 |
"step": 1250
|
85 |
}
|
86 |
],
|
|
|
101 |
"attributes": {}
|
102 |
}
|
103 |
},
|
104 |
+
"total_flos": 1.0191448265781658e+17,
|
105 |
"train_batch_size": 1,
|
106 |
"trial_name": null,
|
107 |
"trial_params": null
|
checkpoint-1250/training_args.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 5560
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:944c7c22023831a73ee4b0a66805723ecef65f25064cb419b24a8d84b3daee22
|
3 |
size 5560
|