Hugging Face's logo

r1char9
/
T5_chat

Text2Text Generation
Transformers
Safetensors
Russian
t5
T5
russian
text-generation
text-generation-inference
Inference Endpoints
Model card Files Community
r1char9 commited on
Commit
a648334
·
verified ·
1 Parent(s): 18f915e

Update config.json

Browse files
Files changed (1) hide show
  1. config.json +62 -62
config.json CHANGED
@@ -1,62 +1,62 @@
1
- {
2
- "_name_or_path": "AlanRobotics/instruct-T5",
3
- "_num_labels": 2,
4
- "architectures": [
5
- "T5ForConditionalGeneration"
6
- ],
7
- "classifier_dropout": 0.0,
8
- "d_ff": 3072,
9
- "d_kv": 64,
10
- "d_model": 768,
11
- "decoder_start_token_id": 0,
12
- "dense_act_fn": "relu",
13
- "dropout_rate": 0.1,
14
- "eos_token_id": 2,
15
- "feed_forward_proj": "relu",
16
- "initializer_factor": 1.0,
17
- "is_encoder_decoder": true,
18
- "is_gated_act": false,
19
- "layer_norm_epsilon": 1e-06,
20
- "model_type": "t5",
21
- "n_positions": 512,
22
- "num_decoder_layers": 12,
23
- "num_heads": 12,
24
- "num_layers": 12,
25
- "output_past": true,
26
- "pad_token_id": 0,
27
- "relative_attention_max_distance": 128,
28
- "relative_attention_num_buckets": 32,
29
- "task_specific_params": {
30
- "summarization": {
31
- "early_stopping": true,
32
- "length_penalty": 2.0,
33
- "max_length": 200,
34
- "min_length": 30,
35
- "no_repeat_ngram_size": 3,
36
- "num_beams": 4,
37
- "prefix": "summarize: "
38
- },
39
- "translation_en_to_de": {
40
- "early_stopping": true,
41
- "max_length": 300,
42
- "num_beams": 4,
43
- "prefix": "translate English to German: "
44
- },
45
- "translation_en_to_fr": {
46
- "early_stopping": true,
47
- "max_length": 300,
48
- "num_beams": 4,
49
- "prefix": "translate English to French: "
50
- },
51
- "translation_en_to_ro": {
52
- "early_stopping": true,
53
- "max_length": 300,
54
- "num_beams": 4,
55
- "prefix": "translate English to Romanian: "
56
- }
57
- },
58
- "torch_dtype": "float32",
59
- "transformers_version": "4.38.2",
60
- "use_cache": true,
61
- "vocab_size": 32128
62
- }
 
1
+ {
2
+ "_name_or_path": "r1char9/T5-chat",
3
+ "_num_labels": 2,
4
+ "architectures": [
5
+ "T5ForConditionalGeneration"
6
+ ],
7
+ "classifier_dropout": 0.0,
8
+ "d_ff": 3072,
9
+ "d_kv": 64,
10
+ "d_model": 768,
11
+ "decoder_start_token_id": 0,
12
+ "dense_act_fn": "relu",
13
+ "dropout_rate": 0.1,
14
+ "eos_token_id": 2,
15
+ "feed_forward_proj": "relu",
16
+ "initializer_factor": 1.0,
17
+ "is_encoder_decoder": true,
18
+ "is_gated_act": false,
19
+ "layer_norm_epsilon": 1e-06,
20
+ "model_type": "t5",
21
+ "n_positions": 512,
22
+ "num_decoder_layers": 12,
23
+ "num_heads": 12,
24
+ "num_layers": 12,
25
+ "output_past": true,
26
+ "pad_token_id": 0,
27
+ "relative_attention_max_distance": 128,
28
+ "relative_attention_num_buckets": 32,
29
+ "task_specific_params": {
30
+ "summarization": {
31
+ "early_stopping": true,
32
+ "length_penalty": 2.0,
33
+ "max_length": 200,
34
+ "min_length": 30,
35
+ "no_repeat_ngram_size": 3,
36
+ "num_beams": 4,
37
+ "prefix": "summarize: "
38
+ },
39
+ "translation_en_to_de": {
40
+ "early_stopping": true,
41
+ "max_length": 300,
42
+ "num_beams": 4,
43
+ "prefix": "translate English to German: "
44
+ },
45
+ "translation_en_to_fr": {
46
+ "early_stopping": true,
47
+ "max_length": 300,
48
+ "num_beams": 4,
49
+ "prefix": "translate English to French: "
50
+ },
51
+ "translation_en_to_ro": {
52
+ "early_stopping": true,
53
+ "max_length": 300,
54
+ "num_beams": 4,
55
+ "prefix": "translate English to Romanian: "
56
+ }
57
+ },
58
+ "torch_dtype": "float32",
59
+ "transformers_version": "4.38.2",
60
+ "use_cache": true,
61
+ "vocab_size": 32128
62
+ }