conjuring92
commited on
Commit
•
4a85089
1
Parent(s):
98a41c2
Training in progress, step 500
Browse files- .gitignore +1 -0
- config.json +27 -0
- pytorch_model.bin +3 -0
- runs/Jan31_21-16-59_jigsaw-rb-compute/1643664148.1373641/events.out.tfevents.1643664148.jigsaw-rb-compute.2721.1 +3 -0
- runs/Jan31_21-16-59_jigsaw-rb-compute/events.out.tfevents.1643664074.jigsaw-rb-compute.2721.0 +3 -0
- runs/Jan31_21-29-07_jigsaw-rb-compute/1643664560.4794006/events.out.tfevents.1643664560.jigsaw-rb-compute.3770.1 +3 -0
- runs/Jan31_21-29-07_jigsaw-rb-compute/events.out.tfevents.1643664560.jigsaw-rb-compute.3770.0 +3 -0
- training_args.bin +3 -0
.gitignore
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
checkpoint-*/
|
config.json
ADDED
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "distilroberta-base",
|
3 |
+
"architectures": [
|
4 |
+
"RobertaForMaskedLM"
|
5 |
+
],
|
6 |
+
"attention_probs_dropout_prob": 0.1,
|
7 |
+
"bos_token_id": 0,
|
8 |
+
"classifier_dropout": null,
|
9 |
+
"eos_token_id": 2,
|
10 |
+
"hidden_act": "gelu",
|
11 |
+
"hidden_dropout_prob": 0.1,
|
12 |
+
"hidden_size": 768,
|
13 |
+
"initializer_range": 0.02,
|
14 |
+
"intermediate_size": 3072,
|
15 |
+
"layer_norm_eps": 1e-05,
|
16 |
+
"max_position_embeddings": 514,
|
17 |
+
"model_type": "roberta",
|
18 |
+
"num_attention_heads": 12,
|
19 |
+
"num_hidden_layers": 6,
|
20 |
+
"pad_token_id": 1,
|
21 |
+
"position_embedding_type": "absolute",
|
22 |
+
"torch_dtype": "float32",
|
23 |
+
"transformers_version": "4.16.0",
|
24 |
+
"type_vocab_size": 1,
|
25 |
+
"use_cache": true,
|
26 |
+
"vocab_size": 50265
|
27 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:78d6171787fc3cccdac4342e484f5309d26daeb94fa289b6a41a428c37313800
|
3 |
+
size 328727211
|
runs/Jan31_21-16-59_jigsaw-rb-compute/1643664148.1373641/events.out.tfevents.1643664148.jigsaw-rb-compute.2721.1
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b85b6b1c78b6f0c812441a68eb116d26858fa4514c32351684323364b8868edd
|
3 |
+
size 4824
|
runs/Jan31_21-16-59_jigsaw-rb-compute/events.out.tfevents.1643664074.jigsaw-rb-compute.2721.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8d0955ee9dd9c4a649d6c5beb1e74ebbcce3cfabdf0f5bfc7e0b3f58e9125fbf
|
3 |
+
size 3584
|
runs/Jan31_21-29-07_jigsaw-rb-compute/1643664560.4794006/events.out.tfevents.1643664560.jigsaw-rb-compute.3770.1
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8b8cb51df173a7a05ee4fd82cecdbba27d18ef75282ecfc698ddaaebc00ec7e9
|
3 |
+
size 4815
|
runs/Jan31_21-29-07_jigsaw-rb-compute/events.out.tfevents.1643664560.jigsaw-rb-compute.3770.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f8fc73511a26b9143cf1193533274c0d79cf0400cde439903a795d5ce63fdf19
|
3 |
+
size 3999
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:00da6e9dc22261da0f1556e6a071fe5ef461c3c991ffacecd7a8bfb2daa7924e
|
3 |
+
size 3055
|