Training in progress, step 500
Browse files- .gitignore +1 -0
- config.json +46 -0
- pytorch_model.bin +3 -0
- runs/Jun03_18-56-13_1fff029d8c40/1685818577.517517/events.out.tfevents.1685818577.1fff029d8c40.1561.1 +3 -0
- runs/Jun03_18-56-13_1fff029d8c40/events.out.tfevents.1685818577.1fff029d8c40.1561.0 +3 -0
- runs/Jun03_18-57-00_1fff029d8c40/1685818623.8560824/events.out.tfevents.1685818623.1fff029d8c40.1561.3 +3 -0
- runs/Jun03_18-57-00_1fff029d8c40/events.out.tfevents.1685818623.1fff029d8c40.1561.2 +3 -0
- runs/Jun03_18-58-58_1fff029d8c40/1685818741.0918477/events.out.tfevents.1685818741.1fff029d8c40.1561.5 +3 -0
- runs/Jun03_18-58-58_1fff029d8c40/events.out.tfevents.1685818741.1fff029d8c40.1561.4 +3 -0
- runs/Jun03_19-00-58_1fff029d8c40/1685818861.132355/events.out.tfevents.1685818861.1fff029d8c40.1561.7 +3 -0
- runs/Jun03_19-00-58_1fff029d8c40/events.out.tfevents.1685818861.1fff029d8c40.1561.6 +3 -0
- runs/Jun03_19-01-46_1fff029d8c40/1685818909.2280815/events.out.tfevents.1685818909.1fff029d8c40.1561.9 +3 -0
- runs/Jun03_19-01-46_1fff029d8c40/events.out.tfevents.1685818909.1fff029d8c40.1561.8 +3 -0
- training_args.bin +3 -0
.gitignore
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
checkpoint-*/
|
config.json
ADDED
@@ -0,0 +1,46 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "distilgpt2",
|
3 |
+
"_num_labels": 1,
|
4 |
+
"activation_function": "gelu_new",
|
5 |
+
"architectures": [
|
6 |
+
"GPT2LMHeadModel"
|
7 |
+
],
|
8 |
+
"attn_pdrop": 0.1,
|
9 |
+
"bos_token_id": 50256,
|
10 |
+
"embd_pdrop": 0.1,
|
11 |
+
"eos_token_id": 50256,
|
12 |
+
"id2label": {
|
13 |
+
"0": "LABEL_0"
|
14 |
+
},
|
15 |
+
"initializer_range": 0.02,
|
16 |
+
"label2id": {
|
17 |
+
"LABEL_0": 0
|
18 |
+
},
|
19 |
+
"layer_norm_epsilon": 1e-05,
|
20 |
+
"model_type": "gpt2",
|
21 |
+
"n_ctx": 1024,
|
22 |
+
"n_embd": 768,
|
23 |
+
"n_head": 12,
|
24 |
+
"n_inner": null,
|
25 |
+
"n_layer": 6,
|
26 |
+
"n_positions": 1024,
|
27 |
+
"reorder_and_upcast_attn": false,
|
28 |
+
"resid_pdrop": 0.1,
|
29 |
+
"scale_attn_by_inverse_layer_idx": false,
|
30 |
+
"scale_attn_weights": true,
|
31 |
+
"summary_activation": null,
|
32 |
+
"summary_first_dropout": 0.1,
|
33 |
+
"summary_proj_to_labels": true,
|
34 |
+
"summary_type": "cls_index",
|
35 |
+
"summary_use_proj": true,
|
36 |
+
"task_specific_params": {
|
37 |
+
"text-generation": {
|
38 |
+
"do_sample": true,
|
39 |
+
"max_length": 50
|
40 |
+
}
|
41 |
+
},
|
42 |
+
"torch_dtype": "float32",
|
43 |
+
"transformers_version": "4.28.0",
|
44 |
+
"use_cache": true,
|
45 |
+
"vocab_size": 50257
|
46 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:824eaa770700efc29052d668a774031b7274f94594cdf5b6e9d06c40959cd477
|
3 |
+
size 333970169
|
runs/Jun03_18-56-13_1fff029d8c40/1685818577.517517/events.out.tfevents.1685818577.1fff029d8c40.1561.1
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:53fa495edbf2dece95d4776d6526d602ac983af77649d26a64a092175c94ad2d
|
3 |
+
size 5900
|
runs/Jun03_18-56-13_1fff029d8c40/events.out.tfevents.1685818577.1fff029d8c40.1561.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0297fd687ce191a3db089772a24f95352785a7ade21c19cddd8761c8cd64d6fa
|
3 |
+
size 4361
|
runs/Jun03_18-57-00_1fff029d8c40/1685818623.8560824/events.out.tfevents.1685818623.1fff029d8c40.1561.3
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8fc3698a14217993baa1bfe82234664a8422b81b600f320f5aeccf2d5669536f
|
3 |
+
size 5900
|
runs/Jun03_18-57-00_1fff029d8c40/events.out.tfevents.1685818623.1fff029d8c40.1561.2
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c15e861bed6ef4b9e649982f30938796b6f6a83b869456d10fce576d6fc25b1a
|
3 |
+
size 4361
|
runs/Jun03_18-58-58_1fff029d8c40/1685818741.0918477/events.out.tfevents.1685818741.1fff029d8c40.1561.5
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:bd4f623640f26f882d5df8b1d7ebff37dc69657baf9069ac97dd297205e3354d
|
3 |
+
size 5900
|
runs/Jun03_18-58-58_1fff029d8c40/events.out.tfevents.1685818741.1fff029d8c40.1561.4
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f9472c6fe73e70dbf695b889bf71b2585b56524802ade806ac626b243166a2fd
|
3 |
+
size 4361
|
runs/Jun03_19-00-58_1fff029d8c40/1685818861.132355/events.out.tfevents.1685818861.1fff029d8c40.1561.7
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2f51c676879c24ded6492fa13201d014b5b8edda3b41e794cd1cce75fb739e65
|
3 |
+
size 5900
|
runs/Jun03_19-00-58_1fff029d8c40/events.out.tfevents.1685818861.1fff029d8c40.1561.6
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:254fa47a991a9a9f63160049874823b897f8e5c8659c43d1f127e198ee56102d
|
3 |
+
size 4360
|
runs/Jun03_19-01-46_1fff029d8c40/1685818909.2280815/events.out.tfevents.1685818909.1fff029d8c40.1561.9
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8c57f3c55996e2445a92087e27f88913560841064729756dc39c1be2b167ca04
|
3 |
+
size 5900
|
runs/Jun03_19-01-46_1fff029d8c40/events.out.tfevents.1685818909.1fff029d8c40.1561.8
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:bd4069480ad9fa897aa299dfe2fb08f708711dec96c42bfd2850b082c16fabb6
|
3 |
+
size 4518
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:40a19c760bd96f50fc47a1bc91e7f9b45ade7f8a7ef9897da143be409cf8cc1e
|
3 |
+
size 3579
|