Upload LEDForConditionalGeneration
Browse files- config.json +157 -0
- generation_config.json +8 -0
- pytorch_model.bin +3 -0
config.json
ADDED
@@ -0,0 +1,157 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "allenai/led-base-16384",
|
3 |
+
"accelerator": null,
|
4 |
+
"accumulate_grad_batches": 2,
|
5 |
+
"activation_dropout": 0.0,
|
6 |
+
"activation_function": "gelu",
|
7 |
+
"add_context": false,
|
8 |
+
"amp_backend": null,
|
9 |
+
"amp_level": null,
|
10 |
+
"architectures": [
|
11 |
+
"LEDForConditionalGeneration"
|
12 |
+
],
|
13 |
+
"attention_dropout": 0.0,
|
14 |
+
"attention_window": [
|
15 |
+
1024,
|
16 |
+
1024,
|
17 |
+
1024,
|
18 |
+
1024,
|
19 |
+
1024,
|
20 |
+
1024
|
21 |
+
],
|
22 |
+
"auto_lr_find": false,
|
23 |
+
"auto_scale_batch_size": false,
|
24 |
+
"auto_select_gpus": null,
|
25 |
+
"batch_size": 8,
|
26 |
+
"benchmark": null,
|
27 |
+
"bos_token_id": 0,
|
28 |
+
"check_val_every_n_epoch": 1,
|
29 |
+
"checkpoint": null,
|
30 |
+
"class_labels": {
|
31 |
+
"delete": 4,
|
32 |
+
"ignore": 0,
|
33 |
+
"merge": 3,
|
34 |
+
"rephrase": 1,
|
35 |
+
"split": 2
|
36 |
+
},
|
37 |
+
"classif_dropout": 0.0,
|
38 |
+
"classifier_dropout": 0.0,
|
39 |
+
"context_dir": null,
|
40 |
+
"context_doc_id": null,
|
41 |
+
"context_window": 13,
|
42 |
+
"d_model": 768,
|
43 |
+
"decoder_attention_heads": 12,
|
44 |
+
"decoder_ffn_dim": 3072,
|
45 |
+
"decoder_layerdrop": 0.0,
|
46 |
+
"decoder_layers": 6,
|
47 |
+
"decoder_start_token_id": 2,
|
48 |
+
"default_root_dir": null,
|
49 |
+
"detect_anomaly": false,
|
50 |
+
"devices": "2",
|
51 |
+
"dropout": 0.1,
|
52 |
+
"enable_checkpointing": true,
|
53 |
+
"enable_model_summary": true,
|
54 |
+
"enable_progress_bar": true,
|
55 |
+
"encoder_attention_heads": 12,
|
56 |
+
"encoder_ffn_dim": 3072,
|
57 |
+
"encoder_layerdrop": 0.0,
|
58 |
+
"encoder_layers": 6,
|
59 |
+
"eos_token_id": 2,
|
60 |
+
"eval_beams": null,
|
61 |
+
"eval_max_length": null,
|
62 |
+
"fast_dev_run": false,
|
63 |
+
"gpus": null,
|
64 |
+
"gradient_checkpointing": false,
|
65 |
+
"gradient_clip_algorithm": null,
|
66 |
+
"gradient_clip_val": null,
|
67 |
+
"id2label": {
|
68 |
+
"0": "LABEL_0",
|
69 |
+
"1": "LABEL_1",
|
70 |
+
"2": "LABEL_2"
|
71 |
+
},
|
72 |
+
"inference_mode": true,
|
73 |
+
"init_std": 0.02,
|
74 |
+
"ipus": null,
|
75 |
+
"is_encoder_decoder": true,
|
76 |
+
"label2id": {
|
77 |
+
"LABEL_0": 0,
|
78 |
+
"LABEL_1": 1,
|
79 |
+
"LABEL_2": 2
|
80 |
+
},
|
81 |
+
"limit_predict_batches": null,
|
82 |
+
"limit_test_batches": null,
|
83 |
+
"limit_train_batches": null,
|
84 |
+
"limit_val_batches": null,
|
85 |
+
"log_every_n_steps": 50,
|
86 |
+
"logger": true,
|
87 |
+
"longformer": true,
|
88 |
+
"lr": 2e-05,
|
89 |
+
"lr_scheduler": false,
|
90 |
+
"max_decoder_position_embeddings": 1024,
|
91 |
+
"max_encoder_position_embeddings": 16384,
|
92 |
+
"max_epochs": null,
|
93 |
+
"max_length": 1024,
|
94 |
+
"max_samples": -1,
|
95 |
+
"max_steps": -1,
|
96 |
+
"max_time": null,
|
97 |
+
"min_epochs": null,
|
98 |
+
"min_steps": null,
|
99 |
+
"model_type": "led",
|
100 |
+
"move_metrics_to_cpu": false,
|
101 |
+
"multiple_trainloader_mode": "max_size_cycle",
|
102 |
+
"name": "o-leddoc",
|
103 |
+
"num_hidden_layers": 6,
|
104 |
+
"num_nodes": 1,
|
105 |
+
"num_processes": null,
|
106 |
+
"num_sanity_val_steps": 2,
|
107 |
+
"op_col": "label",
|
108 |
+
"op_tokens": [
|
109 |
+
"<COPY>",
|
110 |
+
"<REPHRASE>",
|
111 |
+
"<SPLIT>",
|
112 |
+
"<MERGE>",
|
113 |
+
"<DELETE>"
|
114 |
+
],
|
115 |
+
"overfit_batches": 0.0,
|
116 |
+
"pad_token_id": 1,
|
117 |
+
"plan_col": "labels",
|
118 |
+
"plan_loss": null,
|
119 |
+
"plan_prefix": false,
|
120 |
+
"plan_sep": false,
|
121 |
+
"plugins": null,
|
122 |
+
"precision": 32,
|
123 |
+
"prefix_only": false,
|
124 |
+
"profiler": null,
|
125 |
+
"project": "simplification",
|
126 |
+
"reading_lvl": null,
|
127 |
+
"reload_dataloaders_every_n_epochs": 0,
|
128 |
+
"replace_sampler_ddp": true,
|
129 |
+
"resume_from_checkpoint": null,
|
130 |
+
"save_dir": null,
|
131 |
+
"sent_level": false,
|
132 |
+
"simple_context_dir": null,
|
133 |
+
"simple_context_doc_id": null,
|
134 |
+
"skip_val_gen": true,
|
135 |
+
"strategy": null,
|
136 |
+
"sync_batchnorm": false,
|
137 |
+
"torch_dtype": "float32",
|
138 |
+
"tpu_cores": null,
|
139 |
+
"track_grad_norm": -1,
|
140 |
+
"train_check_interval": 0.2,
|
141 |
+
"train_data_dir": null,
|
142 |
+
"train_file": "merge/cochrane_docs_train.csv",
|
143 |
+
"train_split": 0.9,
|
144 |
+
"train_workers": 8,
|
145 |
+
"transformers_version": "4.29.1",
|
146 |
+
"use_cache": true,
|
147 |
+
"val_check_interval": null,
|
148 |
+
"val_file": "merge/cochrane_docs_val.csv",
|
149 |
+
"val_metric": "bleu",
|
150 |
+
"val_split": 0.05,
|
151 |
+
"val_workers": 8,
|
152 |
+
"valid_data_dir": null,
|
153 |
+
"vocab_size": 50275,
|
154 |
+
"wandb_id": null,
|
155 |
+
"x_col": "complex",
|
156 |
+
"y_col": "simple"
|
157 |
+
}
|
generation_config.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"bos_token_id": 0,
|
4 |
+
"decoder_start_token_id": 2,
|
5 |
+
"eos_token_id": 2,
|
6 |
+
"pad_token_id": 1,
|
7 |
+
"transformers_version": "4.29.1"
|
8 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:62bd1dc49161cd4536e74e1f575c88ac3fafc01cfd66b73ce5bfc90b4a61e322
|
3 |
+
size 647711597
|