Update config.json
Browse files- config.json +34 -1
config.json
CHANGED
@@ -5,6 +5,7 @@
|
|
5 |
"d_ff": 16384,
|
6 |
"d_kv": 128,
|
7 |
"d_model": 1024,
|
|
|
8 |
"dropout_rate": 0.1,
|
9 |
"finetuning_task": null,
|
10 |
"initializer_factor": 1.0,
|
@@ -19,7 +20,39 @@
|
|
19 |
"output_past": true,
|
20 |
"pruned_heads": {},
|
21 |
"relative_attention_num_buckets": 32,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
22 |
"torchscript": false,
|
23 |
"use_bfloat16": false,
|
24 |
"vocab_size": 32128
|
25 |
-
}
|
|
|
5 |
"d_ff": 16384,
|
6 |
"d_kv": 128,
|
7 |
"d_model": 1024,
|
8 |
+
"decoder_start_token_id": 0,
|
9 |
"dropout_rate": 0.1,
|
10 |
"finetuning_task": null,
|
11 |
"initializer_factor": 1.0,
|
|
|
20 |
"output_past": true,
|
21 |
"pruned_heads": {},
|
22 |
"relative_attention_num_buckets": 32,
|
23 |
+
"task_specific_params": {
|
24 |
+
"summarization": {
|
25 |
+
"early_stopping": true,
|
26 |
+
"length_penalty": 2.0,
|
27 |
+
"max_length": 200,
|
28 |
+
"min_length": 30,
|
29 |
+
"no_repeat_ngram_size": 3,
|
30 |
+
"num_beams": 4,
|
31 |
+
"prefix": "summarize: "
|
32 |
+
},
|
33 |
+
"translation_en_to_de": {
|
34 |
+
"early_stopping": true,
|
35 |
+
"length_penalty": 2.0,
|
36 |
+
"max_length": 300,
|
37 |
+
"num_beams": 4,
|
38 |
+
"prefix": "translate English to German: "
|
39 |
+
},
|
40 |
+
"translation_en_to_fr": {
|
41 |
+
"early_stopping": true,
|
42 |
+
"length_penalty": 2.0,
|
43 |
+
"max_length": 300,
|
44 |
+
"num_beams": 4,
|
45 |
+
"prefix": "translate English to French: "
|
46 |
+
},
|
47 |
+
"translation_en_to_ro": {
|
48 |
+
"early_stopping": true,
|
49 |
+
"length_penalty": 2.0,
|
50 |
+
"max_length": 300,
|
51 |
+
"num_beams": 4,
|
52 |
+
"prefix": "translate English to Romanian: "
|
53 |
+
}
|
54 |
+
},
|
55 |
"torchscript": false,
|
56 |
"use_bfloat16": false,
|
57 |
"vocab_size": 32128
|
58 |
+
}
|