Commit
•
41efbb1
1
Parent(s):
d424727
Upload 4 files
Browse files
FinalModelcodeT5/config.json
ADDED
@@ -0,0 +1,68 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "Salesforce/codet5-small",
|
3 |
+
"architectures": [
|
4 |
+
"T5ForConditionalGeneration"
|
5 |
+
],
|
6 |
+
"bos_token_id": 1,
|
7 |
+
"d_ff": 2048,
|
8 |
+
"d_kv": 64,
|
9 |
+
"d_model": 512,
|
10 |
+
"decoder_start_token_id": 0,
|
11 |
+
"dense_act_fn": "relu",
|
12 |
+
"dropout_rate": 0.1,
|
13 |
+
"eos_token_id": 2,
|
14 |
+
"feed_forward_proj": "relu",
|
15 |
+
"gradient_checkpointing": false,
|
16 |
+
"id2label": {
|
17 |
+
"0": "LABEL_0"
|
18 |
+
},
|
19 |
+
"initializer_factor": 1.0,
|
20 |
+
"is_encoder_decoder": true,
|
21 |
+
"is_gated_act": false,
|
22 |
+
"label2id": {
|
23 |
+
"LABEL_0": 0
|
24 |
+
},
|
25 |
+
"layer_norm_epsilon": 1e-06,
|
26 |
+
"model_type": "t5",
|
27 |
+
"n_positions": 512,
|
28 |
+
"num_decoder_layers": 6,
|
29 |
+
"num_heads": 8,
|
30 |
+
"num_layers": 6,
|
31 |
+
"output_past": true,
|
32 |
+
"pad_token_id": 0,
|
33 |
+
"relative_attention_max_distance": 128,
|
34 |
+
"relative_attention_num_buckets": 32,
|
35 |
+
"task_specific_params": {
|
36 |
+
"summarization": {
|
37 |
+
"early_stopping": true,
|
38 |
+
"length_penalty": 2.0,
|
39 |
+
"max_length": 200,
|
40 |
+
"min_length": 30,
|
41 |
+
"no_repeat_ngram_size": 3,
|
42 |
+
"num_beams": 4,
|
43 |
+
"prefix": "summarize: "
|
44 |
+
},
|
45 |
+
"translation_en_to_de": {
|
46 |
+
"early_stopping": true,
|
47 |
+
"max_length": 300,
|
48 |
+
"num_beams": 4,
|
49 |
+
"prefix": "translate English to German: "
|
50 |
+
},
|
51 |
+
"translation_en_to_fr": {
|
52 |
+
"early_stopping": true,
|
53 |
+
"max_length": 300,
|
54 |
+
"num_beams": 4,
|
55 |
+
"prefix": "translate English to French: "
|
56 |
+
},
|
57 |
+
"translation_en_to_ro": {
|
58 |
+
"early_stopping": true,
|
59 |
+
"max_length": 300,
|
60 |
+
"num_beams": 4,
|
61 |
+
"prefix": "translate English to Romanian: "
|
62 |
+
}
|
63 |
+
},
|
64 |
+
"torch_dtype": "float32",
|
65 |
+
"transformers_version": "4.21.3",
|
66 |
+
"use_cache": true,
|
67 |
+
"vocab_size": 32100
|
68 |
+
}
|
FinalModelcodeT5/pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:50c3369f26aa206d62161f73cc5312018491beb373e3f6e5fd6d79981dd72def
|
3 |
+
size 242011067
|
PLBART_FinalModelDir/config.json
ADDED
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "uclanlp/plbart-base",
|
3 |
+
"activation_dropout": 0.0,
|
4 |
+
"activation_function": "gelu",
|
5 |
+
"architectures": [
|
6 |
+
"PLBartForConditionalGeneration"
|
7 |
+
],
|
8 |
+
"attention_dropout": 0.1,
|
9 |
+
"bos_token_id": 0,
|
10 |
+
"classifier_dropout": 0.0,
|
11 |
+
"d_model": 768,
|
12 |
+
"decoder_attention_heads": 12,
|
13 |
+
"decoder_ffn_dim": 3072,
|
14 |
+
"decoder_layerdrop": 0.0,
|
15 |
+
"decoder_layers": 6,
|
16 |
+
"dropout": 0.1,
|
17 |
+
"encoder_attention_heads": 12,
|
18 |
+
"encoder_ffn_dim": 3072,
|
19 |
+
"encoder_layerdrop": 0.0,
|
20 |
+
"encoder_layers": 6,
|
21 |
+
"eos_token_id": 2,
|
22 |
+
"forced_eos_token_id": 2,
|
23 |
+
"init_std": 0.02,
|
24 |
+
"is_encoder_decoder": true,
|
25 |
+
"max_position_embeddings": 1024,
|
26 |
+
"model_type": "plbart",
|
27 |
+
"num_hidden_layers": 6,
|
28 |
+
"pad_token_id": 1,
|
29 |
+
"scale_embedding": true,
|
30 |
+
"torch_dtype": "float32",
|
31 |
+
"transformers_version": "4.22.1",
|
32 |
+
"use_cache": true,
|
33 |
+
"vocab_size": 50005
|
34 |
+
}
|
PLBART_FinalModelDir/pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3e30f658583d027817214f96f3d50e1225f66b9fe8b6d932019131860929f1ce
|
3 |
+
size 557165689
|