Update config.json
Browse files- config.json +21 -2
config.json
CHANGED
|
@@ -1,15 +1,20 @@
|
|
| 1 |
{
|
|
|
|
| 2 |
"architectures": [
|
| 3 |
"FlaubertWithLMHeadModel"
|
| 4 |
],
|
| 5 |
-
"
|
| 6 |
"attention_dropout": 0.1,
|
| 7 |
"bos_index": 0,
|
|
|
|
| 8 |
"bptt": 512,
|
|
|
|
| 9 |
"clip_grad_norm": 5,
|
| 10 |
"dropout": 0.1,
|
| 11 |
"emb_dim": 512,
|
|
|
|
| 12 |
"encoder_only": true,
|
|
|
|
| 13 |
"eos_index": 1,
|
| 14 |
"fp16": true,
|
| 15 |
"gelu_activation": true,
|
|
@@ -17,9 +22,12 @@
|
|
| 17 |
"id2lang": {
|
| 18 |
"0": "fr"
|
| 19 |
},
|
|
|
|
|
|
|
| 20 |
"lang2id": {
|
| 21 |
"fr": 0
|
| 22 |
},
|
|
|
|
| 23 |
"langs": [
|
| 24 |
"fr"
|
| 25 |
],
|
|
@@ -28,7 +36,9 @@
|
|
| 28 |
"lg_sampling_factor": -1,
|
| 29 |
"lgs": "fr",
|
| 30 |
"mask_index": 5,
|
|
|
|
| 31 |
"max_batch_size": 0,
|
|
|
|
| 32 |
"max_vocab": -1,
|
| 33 |
"mlm_steps": [
|
| 34 |
[
|
|
@@ -36,17 +46,26 @@
|
|
| 36 |
null
|
| 37 |
]
|
| 38 |
],
|
|
|
|
| 39 |
"n_heads": 8,
|
| 40 |
"n_langs": 1,
|
| 41 |
"n_layers": 6,
|
| 42 |
-
"n_words": 68729,
|
| 43 |
"pad_index": 2,
|
|
|
|
| 44 |
"pre_norm": true,
|
| 45 |
"sample_alpha": 0,
|
| 46 |
"share_inout_emb": true,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 47 |
"tokens_per_batch": -1,
|
| 48 |
"unk_index": 3,
|
| 49 |
"use_lang_emb": true,
|
|
|
|
| 50 |
"word_blank": 0,
|
| 51 |
"word_dropout": 0,
|
| 52 |
"word_keep": 0.1,
|
|
|
|
| 1 |
{
|
| 2 |
+
"amp": 1,
|
| 3 |
"architectures": [
|
| 4 |
"FlaubertWithLMHeadModel"
|
| 5 |
],
|
| 6 |
+
"asm": false,
|
| 7 |
"attention_dropout": 0.1,
|
| 8 |
"bos_index": 0,
|
| 9 |
+
"bos_token_id": 0,
|
| 10 |
"bptt": 512,
|
| 11 |
+
"causal": false,
|
| 12 |
"clip_grad_norm": 5,
|
| 13 |
"dropout": 0.1,
|
| 14 |
"emb_dim": 512,
|
| 15 |
+
"embed_init_std": 0.02209708691207961,
|
| 16 |
"encoder_only": true,
|
| 17 |
+
"end_n_top": 5,
|
| 18 |
"eos_index": 1,
|
| 19 |
"fp16": true,
|
| 20 |
"gelu_activation": true,
|
|
|
|
| 22 |
"id2lang": {
|
| 23 |
"0": "fr"
|
| 24 |
},
|
| 25 |
+
"init_std": 0.02,
|
| 26 |
+
"is_encoder": true,
|
| 27 |
"lang2id": {
|
| 28 |
"fr": 0
|
| 29 |
},
|
| 30 |
+
"lang_id": 0,
|
| 31 |
"langs": [
|
| 32 |
"fr"
|
| 33 |
],
|
|
|
|
| 36 |
"lg_sampling_factor": -1,
|
| 37 |
"lgs": "fr",
|
| 38 |
"mask_index": 5,
|
| 39 |
+
"mask_token_id": 0,
|
| 40 |
"max_batch_size": 0,
|
| 41 |
+
"max_position_embeddings": 512,
|
| 42 |
"max_vocab": -1,
|
| 43 |
"mlm_steps": [
|
| 44 |
[
|
|
|
|
| 46 |
null
|
| 47 |
]
|
| 48 |
],
|
| 49 |
+
"model_type": "flaubert",
|
| 50 |
"n_heads": 8,
|
| 51 |
"n_langs": 1,
|
| 52 |
"n_layers": 6,
|
|
|
|
| 53 |
"pad_index": 2,
|
| 54 |
+
"pad_token_id": 2,
|
| 55 |
"pre_norm": true,
|
| 56 |
"sample_alpha": 0,
|
| 57 |
"share_inout_emb": true,
|
| 58 |
+
"sinusoidal_embeddings": false,
|
| 59 |
+
"start_n_top": 5,
|
| 60 |
+
"summary_activation": null,
|
| 61 |
+
"summary_first_dropout": 0.1,
|
| 62 |
+
"summary_proj_to_labels": true,
|
| 63 |
+
"summary_type": "first",
|
| 64 |
+
"summary_use_proj": true,
|
| 65 |
"tokens_per_batch": -1,
|
| 66 |
"unk_index": 3,
|
| 67 |
"use_lang_emb": true,
|
| 68 |
+
"vocab_size": 68729,
|
| 69 |
"word_blank": 0,
|
| 70 |
"word_dropout": 0,
|
| 71 |
"word_keep": 0.1,
|