Daniel Stancl
commited on
Commit
•
19164a5
1
Parent(s):
48f3a5e
Add config file and PT+Flax ckpts
Browse files- config.json +32 -0
- flax_model.msgpack +3 -0
- pytorch_model.bin +3 -0
config.json
ADDED
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "Stancld/LongT5-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 |
+
"dropout_rate": 0.1,
|
11 |
+
"encoder_attention_type": "transient-global",
|
12 |
+
"eos_token_id": 1,
|
13 |
+
"feed_forward_proj": "gated-gelu",
|
14 |
+
"global_block_size": 16,
|
15 |
+
"initializer_factor": 1.0,
|
16 |
+
"is_encoder_decoder": true,
|
17 |
+
"layer_norm_epsilon": 1e-06,
|
18 |
+
"local_radius": 127,
|
19 |
+
"model_type": "longt5",
|
20 |
+
"n_positions": 4096,
|
21 |
+
"num_decoder_layers": 12,
|
22 |
+
"num_heads": 12,
|
23 |
+
"num_layers": 12,
|
24 |
+
"output_past": true,
|
25 |
+
"pad_token_id": 0,
|
26 |
+
"relative_attention_max_distance": 128,
|
27 |
+
"relative_attention_num_buckets": 32,
|
28 |
+
"torch_dtype": "float32",
|
29 |
+
"transformers_version": "4.19.0.dev0",
|
30 |
+
"use_cache": true,
|
31 |
+
"vocab_size": 32128
|
32 |
+
}
|
flax_model.msgpack
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d7ace4230e672bbada0da3dc7dcc38e0dd59adb897e54df1b2ebc01bd63bdfd1
|
3 |
+
size 990362880
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9f898cd5adc25ba354c83bff5c6734bf1947e762a7399eae1094fcf38860e6ce
|
3 |
+
size 891748981
|