update config
Browse files- config.json +8 -4
- pytorch_model.bin +2 -2
config.json
CHANGED
|
@@ -1,14 +1,16 @@
|
|
| 1 |
{
|
|
|
|
| 2 |
"activation_dropout": 0.0,
|
| 3 |
"activation_function": "gelu",
|
| 4 |
"add_bias_logits": false,
|
| 5 |
"add_final_layer_norm": false,
|
| 6 |
"architectures": [
|
| 7 |
-
"
|
| 8 |
],
|
| 9 |
"attention_dropout": 0.0,
|
| 10 |
"bos_token_id": 1,
|
| 11 |
"classif_dropout": 0.0,
|
|
|
|
| 12 |
"d_model": 512,
|
| 13 |
"decoder_attention_heads": 16,
|
| 14 |
"decoder_ffn_dim": 2048,
|
|
@@ -23,6 +25,8 @@
|
|
| 23 |
"eos_token_id": 2,
|
| 24 |
"extra_pos_embeddings": 0,
|
| 25 |
"force_bos_token_to_be_generated": false,
|
|
|
|
|
|
|
| 26 |
"id2label": {
|
| 27 |
"0": "LABEL_0",
|
| 28 |
"1": "LABEL_1",
|
|
@@ -39,7 +43,7 @@
|
|
| 39 |
"max_length": 128,
|
| 40 |
"max_position_embeddings": 512,
|
| 41 |
"min_length": 20,
|
| 42 |
-
"model_type": "blenderbot",
|
| 43 |
"no_repeat_ngram_size": 3,
|
| 44 |
"normalize_before": false,
|
| 45 |
"normalize_embedding": true,
|
|
@@ -48,8 +52,8 @@
|
|
| 48 |
"pad_token_id": 0,
|
| 49 |
"scale_embedding": true,
|
| 50 |
"static_position_embeddings": false,
|
|
|
|
| 51 |
"unk_token_id": 3,
|
| 52 |
-
"
|
| 53 |
-
"do_blenderbot_90_layernorm": true,
|
| 54 |
"vocab_size": 54944
|
| 55 |
}
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "blenderbot-90M/",
|
| 3 |
"activation_dropout": 0.0,
|
| 4 |
"activation_function": "gelu",
|
| 5 |
"add_bias_logits": false,
|
| 6 |
"add_final_layer_norm": false,
|
| 7 |
"architectures": [
|
| 8 |
+
"BlenderbotSmallForConditionalGeneration"
|
| 9 |
],
|
| 10 |
"attention_dropout": 0.0,
|
| 11 |
"bos_token_id": 1,
|
| 12 |
"classif_dropout": 0.0,
|
| 13 |
+
"classifier_dropout": 0.0,
|
| 14 |
"d_model": 512,
|
| 15 |
"decoder_attention_heads": 16,
|
| 16 |
"decoder_ffn_dim": 2048,
|
| 25 |
"eos_token_id": 2,
|
| 26 |
"extra_pos_embeddings": 0,
|
| 27 |
"force_bos_token_to_be_generated": false,
|
| 28 |
+
"forced_eos_token_id": 2,
|
| 29 |
+
"gradient_checkpointing": false,
|
| 30 |
"id2label": {
|
| 31 |
"0": "LABEL_0",
|
| 32 |
"1": "LABEL_1",
|
| 43 |
"max_length": 128,
|
| 44 |
"max_position_embeddings": 512,
|
| 45 |
"min_length": 20,
|
| 46 |
+
"model_type": "blenderbot-small",
|
| 47 |
"no_repeat_ngram_size": 3,
|
| 48 |
"normalize_before": false,
|
| 49 |
"normalize_embedding": true,
|
| 52 |
"pad_token_id": 0,
|
| 53 |
"scale_embedding": true,
|
| 54 |
"static_position_embeddings": false,
|
| 55 |
+
"transformers_version": "4.4.0.dev0",
|
| 56 |
"unk_token_id": 3,
|
| 57 |
+
"use_cache": true,
|
|
|
|
| 58 |
"vocab_size": 54944
|
| 59 |
}
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a6cbcd7e7157bc5ed8cac3e734bb4622852c3932fb8de297cb0b5f87d0990e73
|
| 3 |
+
size 350387079
|