kresnandika
commited on
Commit
•
84ded55
1
Parent(s):
e46add0
Training in progress, epoch 1
Browse files- .gitignore +1 -0
- config.json +35 -0
- logs/1674813019.5581293/events.out.tfevents.1674813019.ip-172-31-27-96.17248.1 +3 -0
- logs/1674821897.4050336/events.out.tfevents.1674821897.ip-172-31-27-96.18529.4 +3 -0
- logs/events.out.tfevents.1674813019.ip-172-31-27-96.17248.0 +3 -0
- logs/events.out.tfevents.1674821897.ip-172-31-27-96.18529.3 +3 -0
- pytorch_model.bin +3 -0
- training_args.bin +3 -0
.gitignore
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
checkpoint-*/
|
config.json
ADDED
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "google/long-t5-tglobal-base",
|
3 |
+
"architectures": [
|
4 |
+
"LongT5ForConditionalGeneration"
|
5 |
+
],
|
6 |
+
"d_ff": 2048,
|
7 |
+
"d_kv": 64,
|
8 |
+
"d_model": 768,
|
9 |
+
"decoder_start_token_id": 0,
|
10 |
+
"dense_act_fn": "gelu_new",
|
11 |
+
"dropout_rate": 0.1,
|
12 |
+
"encoder_attention_type": "transient-global",
|
13 |
+
"eos_token_id": 1,
|
14 |
+
"feed_forward_proj": "gated-gelu",
|
15 |
+
"global_block_size": 16,
|
16 |
+
"initializer_factor": 1.0,
|
17 |
+
"is_encoder_decoder": true,
|
18 |
+
"is_gated_act": true,
|
19 |
+
"layer_norm_epsilon": 1e-06,
|
20 |
+
"local_radius": 127,
|
21 |
+
"model_type": "longt5",
|
22 |
+
"n_positions": 4096,
|
23 |
+
"num_decoder_layers": 12,
|
24 |
+
"num_heads": 12,
|
25 |
+
"num_layers": 12,
|
26 |
+
"output_past": true,
|
27 |
+
"pad_token_id": 0,
|
28 |
+
"relative_attention_max_distance": 128,
|
29 |
+
"relative_attention_num_buckets": 32,
|
30 |
+
"tie_word_embeddings": false,
|
31 |
+
"torch_dtype": "float32",
|
32 |
+
"transformers_version": "4.26.0",
|
33 |
+
"use_cache": true,
|
34 |
+
"vocab_size": 32128
|
35 |
+
}
|
logs/1674813019.5581293/events.out.tfevents.1674813019.ip-172-31-27-96.17248.1
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:59555c018b039d93bf71a3de22c5253b99010905dafcf145224b64f767e21965
|
3 |
+
size 5933
|
logs/1674821897.4050336/events.out.tfevents.1674821897.ip-172-31-27-96.18529.4
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:893a8691526c8f2268308be8bfa3b99726c633900861f0e92a41abef849ddfb4
|
3 |
+
size 5927
|
logs/events.out.tfevents.1674813019.ip-172-31-27-96.17248.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5d56911ea06e2c376818edf0e626738b267baf748c22d3aba401ba67d0b52efd
|
3 |
+
size 4136
|
logs/events.out.tfevents.1674821897.ip-172-31-27-96.18529.3
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:953f0f2d6b3bf93f928a4a4a0152b432fd179dbc3a5ab5db578e3caf4e9a05b1
|
3 |
+
size 6941
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7a75496bf779c8874d101c3e7c06781b070581345bdb987a4ff512d0059898ad
|
3 |
+
size 990452905
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0089d448aabf2fa9337045255cdf6579de593c40610776da4290c38cb763eec1
|
3 |
+
size 3643
|