neuronxcc-2.15.128.0+56dc5a86/0_REGISTRY/0.0.25.dev0/inference/xlm-roberta/hf-internal-testing/tiny-xlm-roberta/fe924f54747c457b124f.json

{"architectures": ["XLMRobertaForCausalLM"], "attention_probs_dropout_prob": 0.1, "classifier_dropout": null, "d_ff": 256, "d_kv": 8, "d_model": 64, "gradient_checkpointing": false, "hidden_act": "gelu", "hidden_dropout_prob": 0.1, "hidden_size": 256, "initializer_range": 0.02, "intermediate_size": 256, "layer_norm_eps": 1e-05, "max_position_embeddings": 64, "model_type": "xlm-roberta", "neuron": {"auto_cast": null, "auto_cast_type": null, "compiler_type": "neuronx-cc", "compiler_version": "2.15.128.0+56dc5a86", "disable_fallback": false, "disable_fast_relayout": false, "dynamic_batch_size": false, "inline_weights_to_neff": true, "optlevel": "2", "output_attentions": false, "output_hidden_states": false, "static_batch_size": 1, "static_sequence_length": 32}, "num_attention_heads": 2, "num_decoder_layers": 2, "num_heads": 2, "num_hidden_layers": 2, "num_layers": 2, "output_past": true, "position_embedding_type": "absolute", "relative_attention_num_buckets": 32, "task": "text-classification", "type_vocab_size": 1, "use_cache": true, "vocab_size": 5002}