Hugging Face's logo

ArthurZ
/
umt5-small

Transformers
PyTorch
mt5
text2text-generation
Model card Files
xet
Community
1
ArthurZ HF Staff commited on
Commit
535e231
·
1 Parent(s): 1ef81ae

Upload MT5ForConditionalGeneration

Browse files
Files changed (2) hide show
  1. config.json +3 -3
  2. pytorch_model.bin +2 -2
config.json CHANGED
@@ -7,10 +7,10 @@
7
  "d_kv": 64,
8
  "d_model": 512,
9
  "decoder_start_token_id": 0,
10
- "dense_act_fn": "silu",
11
  "dropout_rate": 0.1,
12
  "eos_token_id": 1,
13
- "feed_forward_proj": "gated-silu",
14
  "initializer_factor": 1.0,
15
  "is_encoder_decoder": true,
16
  "is_gated_act": true,
@@ -24,7 +24,7 @@
24
  "relative_attention_max_distance": 128,
25
  "relative_attention_num_buckets": 32,
26
  "scalable_attention": true,
27
- "tie_word_embeddings": false,
28
  "tokenizer_class": "T5Tokenizer",
29
  "torch_dtype": "float32",
30
  "transformers_version": "4.31.0.dev0",
 
7
  "d_kv": 64,
8
  "d_model": 512,
9
  "decoder_start_token_id": 0,
10
+ "dense_act_fn": "gelu_new",
11
  "dropout_rate": 0.1,
12
  "eos_token_id": 1,
13
+ "feed_forward_proj": "gated-gelu",
14
  "initializer_factor": 1.0,
15
  "is_encoder_decoder": true,
16
  "is_gated_act": true,
 
24
  "relative_attention_max_distance": 128,
25
  "relative_attention_num_buckets": 32,
26
  "scalable_attention": true,
27
+ "tie_word_embeddings": true,
28
  "tokenizer_class": "T5Tokenizer",
29
  "torch_dtype": "float32",
30
  "transformers_version": "4.31.0.dev0",
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:94ee3bfb9c3b281fadb2dbd3e9d4230c9c6d56cb501d25089435eb5e46552aa0
3
- size 1226475581
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:94c1d78cea32e3b1fabd804f7da1fa591a6e44275c55c9bcf7fd8a7a3d7f899a
3
+ size 701400953