Commit
•
fbbd34d
1
Parent(s):
58368df
Remove custom tokenizer, add model checkpoint
Browse files- config.json +28 -0
- merges.txt +0 -0
- pytorch_model.bin +3 -0
- vocab.json +0 -0
config.json
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_dropout": 0.0,
|
| 3 |
+
"activation_function": "gelu",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"MegatronForCausalLM"
|
| 6 |
+
],
|
| 7 |
+
"attention_dropout": 0.1,
|
| 8 |
+
"bos_token_id": 0,
|
| 9 |
+
"classifier_dropout": 0.0,
|
| 10 |
+
"d_model": 3072,
|
| 11 |
+
"decoder_attention_heads": 32,
|
| 12 |
+
"decoder_ffn_dim": 18432,
|
| 13 |
+
"decoder_layerdrop": 0,
|
| 14 |
+
"decoder_layers": 72,
|
| 15 |
+
"decoder_start_token_id": 0,
|
| 16 |
+
"dropout": 0.1,
|
| 17 |
+
"eos_token_id": 2,
|
| 18 |
+
"gradient_checkpointing": false,
|
| 19 |
+
"init_std": 0.02,
|
| 20 |
+
"is_decoder": true,
|
| 21 |
+
"max_position_embeddings": 1024,
|
| 22 |
+
"model_type": "megatron",
|
| 23 |
+
"pad_token_id": 1,
|
| 24 |
+
"scale_embedding": true,
|
| 25 |
+
"transformers_version": "4.4.0.dev0",
|
| 26 |
+
"use_cache": true,
|
| 27 |
+
"vocab_size": 51200
|
| 28 |
+
}
|
merges.txt
DELETED
|
The diff for this file is too large to render.
See raw diff
|
|
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:09c1cde82a7fe47e0f4d45b1eef6e4d7f23f05271b3214eb1f79c537e9283104
|
| 3 |
+
size 44129367544
|
vocab.json
DELETED
|
The diff for this file is too large to render.
See raw diff
|
|
|