| { |
| "architectures": [ |
| "MarkupLMForPretraining" |
| ], |
| "attention_probs_dropout_prob": 0.1, |
| "bos_token_id": 0, |
| "classifier_dropout": null, |
| "eos_token_id": 2, |
| "gradient_checkpointing": false, |
| "hidden_act": "gelu", |
| "hidden_dropout_prob": 0.1, |
| "hidden_size": 1024, |
| "initializer_range": 0.02, |
| "intermediate_size": 4096, |
| "layer_norm_eps": 1e-05, |
| "max_depth": 50, |
| "max_position_embeddings": 514, |
| "max_xpath_subs_unit_embeddings": 1024, |
| "max_xpath_tag_unit_embeddings": 256, |
| "model_type": "markuplm", |
| "num_attention_heads": 16, |
| "num_hidden_layers": 24, |
| "pad_token_id": 1, |
| "position_embedding_type": "absolute", |
| "torch_dtype": "float16", |
| "transformers_version": "4.10.2", |
| "type_vocab_size": 1, |
| "use_cache": true, |
| "vocab_size": 50267, |
| "xpath_unit_hidden_size": 32, |
| "widget": { |
| "examples": [ |
| { |
| "html": "<html><body><h1>What is MarkupLM?</h1><p>MarkupLM is a model for understanding structured documents.</p></body></html>", |
| "question": "What is MarkupLM?" |
| } |
| ] |
| } |
| } |
|
|