erfan226 commited on
Commit
a0dccdc
1 Parent(s): fe82a33
Files changed (2) hide show
  1. config.json +5 -2
  2. pytorch_model.bin +2 -2
config.json CHANGED
@@ -1,5 +1,5 @@
1
  {
2
- "_name_or_path": "/gd2/MyDrive/models/persian-t5-style-paraphraser",
3
  "architectures": [
4
  "T5ForConditionalGeneration"
5
  ],
@@ -7,11 +7,13 @@
7
  "d_kv": 64,
8
  "d_model": 768,
9
  "decoder_start_token_id": 0,
 
10
  "dropout_rate": 0.1,
11
  "eos_token_id": 1,
12
  "feed_forward_proj": "gated-gelu",
13
  "initializer_factor": 1.0,
14
  "is_encoder_decoder": true,
 
15
  "layer_norm_epsilon": 1e-06,
16
  "model_type": "t5",
17
  "num_decoder_layers": 12,
@@ -19,10 +21,11 @@
19
  "num_layers": 12,
20
  "output_past": true,
21
  "pad_token_id": 0,
 
22
  "relative_attention_num_buckets": 32,
23
  "tie_word_embeddings": false,
24
  "torch_dtype": "float32",
25
- "transformers_version": "4.16.2",
26
  "use_cache": true,
27
  "vocab_size": 32103
28
  }
 
1
  {
2
+ "_name_or_path": "/content/drive/MyDrive/models/st-with-paraphraser-weights-sentinel-token",
3
  "architectures": [
4
  "T5ForConditionalGeneration"
5
  ],
 
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
  "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,
17
  "layer_norm_epsilon": 1e-06,
18
  "model_type": "t5",
19
  "num_decoder_layers": 12,
 
21
  "num_layers": 12,
22
  "output_past": true,
23
  "pad_token_id": 0,
24
+ "relative_attention_max_distance": 128,
25
  "relative_attention_num_buckets": 32,
26
  "tie_word_embeddings": false,
27
  "torch_dtype": "float32",
28
+ "transformers_version": "4.21.2",
29
  "use_cache": true,
30
  "vocab_size": 32103
31
  }
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a745487322912a3b6355e1ebc2c9ba2fbac15971149c6724c284e337278f5d6a
3
- size 990284749
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e37f43599ee6fcf956b9bd61f7fb7893a8694549584472f4ba5234c937d2c71c
3
+ size 990249037