valhalla commited on
Commit
76041a4
1 Parent(s): 22fa338

add flax model

Browse files
Files changed (3) hide show
  1. .gitattributes +1 -0
  2. config.json +11 -8
  3. flax_model.msgpack +3 -0
.gitattributes CHANGED
@@ -6,3 +6,4 @@
6
  *.tar.gz filter=lfs diff=lfs merge=lfs -text
7
  *.ot filter=lfs diff=lfs merge=lfs -text
8
  *.onnx filter=lfs diff=lfs merge=lfs -text
 
 
6
  *.tar.gz filter=lfs diff=lfs merge=lfs -text
7
  *.ot filter=lfs diff=lfs merge=lfs -text
8
  *.onnx filter=lfs diff=lfs merge=lfs -text
9
+ *.msgpack filter=lfs diff=lfs merge=lfs -text
config.json CHANGED
@@ -4,13 +4,12 @@
4
  "add_bias_logits": false,
5
  "add_final_layer_norm": false,
6
  "architectures": [
7
- "BartModel",
8
- "BartForConditionalGeneration",
9
- "BartForSequenceClassification"
10
  ],
11
  "attention_dropout": 0.1,
12
  "bos_token_id": 0,
13
  "classif_dropout": 0.1,
 
14
  "d_model": 1024,
15
  "decoder_attention_heads": 16,
16
  "decoder_ffn_dim": 4096,
@@ -24,6 +23,8 @@
24
  "encoder_layerdrop": 0.0,
25
  "encoder_layers": 12,
26
  "eos_token_id": 2,
 
 
27
  "id2label": {
28
  "0": "LABEL_0",
29
  "1": "LABEL_1",
@@ -38,20 +39,20 @@
38
  },
39
  "max_position_embeddings": 1024,
40
  "model_type": "bart",
 
41
  "normalize_before": false,
 
42
  "num_hidden_layers": 12,
43
  "pad_token_id": 1,
44
  "scale_embedding": false,
45
- "num_beams": 4,
46
- "no_repeat_ngram_size": 3,
47
  "task_specific_params": {
48
- "summarization": {
49
- "length_penalty": 1.0,
50
  "max_length": 128,
51
  "min_length": 12,
52
  "num_beams": 4
53
  },
54
- "summarization_cnn": {
55
  "length_penalty": 2.0,
56
  "max_length": 142,
57
  "min_length": 56,
@@ -64,5 +65,7 @@
64
  "num_beams": 6
65
  }
66
  },
 
 
67
  "vocab_size": 50265
68
  }
 
4
  "add_bias_logits": false,
5
  "add_final_layer_norm": false,
6
  "architectures": [
7
+ "BartModel"
 
 
8
  ],
9
  "attention_dropout": 0.1,
10
  "bos_token_id": 0,
11
  "classif_dropout": 0.1,
12
+ "classifier_dropout": 0.0,
13
  "d_model": 1024,
14
  "decoder_attention_heads": 16,
15
  "decoder_ffn_dim": 4096,
 
23
  "encoder_layerdrop": 0.0,
24
  "encoder_layers": 12,
25
  "eos_token_id": 2,
26
+ "forced_eos_token_id": 2,
27
+ "gradient_checkpointing": false,
28
  "id2label": {
29
  "0": "LABEL_0",
30
  "1": "LABEL_1",
 
39
  },
40
  "max_position_embeddings": 1024,
41
  "model_type": "bart",
42
+ "no_repeat_ngram_size": 3,
43
  "normalize_before": false,
44
+ "num_beams": 4,
45
  "num_hidden_layers": 12,
46
  "pad_token_id": 1,
47
  "scale_embedding": false,
 
 
48
  "task_specific_params": {
49
+ "summarization": {
50
+ "length_penalty": 1.0,
51
  "max_length": 128,
52
  "min_length": 12,
53
  "num_beams": 4
54
  },
55
+ "summarization_cnn": {
56
  "length_penalty": 2.0,
57
  "max_length": 142,
58
  "min_length": 56,
 
65
  "num_beams": 6
66
  }
67
  },
68
+ "transformers_version": "4.7.0.dev0",
69
+ "use_cache": true,
70
  "vocab_size": 50265
71
  }
flax_model.msgpack ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:954a481dd580ea66a4c221ffc66486f2dc6850f8012e16428a4f8d45f67718f9
3
+ size 812600864