Upload 2 files
Browse files- config.json +61 -0
- pytorch_model.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"approx_mode": "full",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"MraForMaskedLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"block_per_row": 4,
|
| 8 |
+
"bos_token_id": 0,
|
| 9 |
+
"dataset": "datasets/roberta-train-sl_512-bs_512-nb_1000-ne_200-bookcorpus-english_wiki",
|
| 10 |
+
"eos_token_id": 2,
|
| 11 |
+
"from_cp": "init.model",
|
| 12 |
+
"from_pretrained_roberta": false,
|
| 13 |
+
"gpu_setting": {
|
| 14 |
+
"inst_per_gpu": 16
|
| 15 |
+
},
|
| 16 |
+
"hidden_act": "gelu",
|
| 17 |
+
"hidden_dropout_prob": 0.1,
|
| 18 |
+
"hidden_size": 768,
|
| 19 |
+
"initial_prior_diagonal_n_blocks": 0,
|
| 20 |
+
"initial_prior_first_n_blocks": 0,
|
| 21 |
+
"initializer_range": 0.02,
|
| 22 |
+
"intermediate_size": 3072,
|
| 23 |
+
"layer_norm_eps": 1e-05,
|
| 24 |
+
"max_position_embeddings": 512,
|
| 25 |
+
"model": {
|
| 26 |
+
"approx_mode": "full",
|
| 27 |
+
"block_per_row": 4,
|
| 28 |
+
"dim": 768,
|
| 29 |
+
"dropout_prob": 0.1,
|
| 30 |
+
"embedding_dim": 768,
|
| 31 |
+
"head_dim": 64,
|
| 32 |
+
"hidden_dim": 3072,
|
| 33 |
+
"initial_prior_diagonal_n_blocks": 0,
|
| 34 |
+
"initial_prior_first_n_blocks": 0,
|
| 35 |
+
"max_seq_len": 512,
|
| 36 |
+
"mixed_precision": true,
|
| 37 |
+
"model_type": "mra2",
|
| 38 |
+
"num_head": 12,
|
| 39 |
+
"num_layers": 12,
|
| 40 |
+
"num_sen_type": 1,
|
| 41 |
+
"shared_weight": false,
|
| 42 |
+
"vocab_size": 50265
|
| 43 |
+
},
|
| 44 |
+
"model_type": "mra",
|
| 45 |
+
"num_attention_heads": 12,
|
| 46 |
+
"num_hidden_layers": 12,
|
| 47 |
+
"pad_token_id": 1,
|
| 48 |
+
"position_embedding_type": "absolute",
|
| 49 |
+
"pretraining_setting": {
|
| 50 |
+
"batches_per_epoch": 1000,
|
| 51 |
+
"batches_per_report": 10,
|
| 52 |
+
"epoch": 20,
|
| 53 |
+
"learning_rate": 5e-05,
|
| 54 |
+
"warmup": 0.05
|
| 55 |
+
},
|
| 56 |
+
"torch_dtype": "float32",
|
| 57 |
+
"transformers_version": "4.31.0.dev0",
|
| 58 |
+
"type_vocab_size": 1,
|
| 59 |
+
"val_dataset": "datasets/roberta-val-sl_512-bs_512-nb_200-ne_1-bookcorpus-english_wiki-transformer/epoch-0000.pickle.gzip",
|
| 60 |
+
"vocab_size": 50265
|
| 61 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e5de1e48b7c21afe66a9abfb4cddcf24ae387ae3ab3ecf20de3a259506e80bf0
|
| 3 |
+
size 498858283
|