| { |
| "builder_config": { |
| "hidden_size": 1280, |
| "int8": false, |
| "max_batch_size": 8, |
| "max_beam_width": 4, |
| "n_mels": 128, |
| "name": "whisper_encoder", |
| "num_heads": 20, |
| "num_languages": 100, |
| "num_layers": 32, |
| "precision": "float16", |
| "strongly_typed": false, |
| "tensor_parallel": 1, |
| "use_refit": false |
| }, |
| "plugin_config": { |
| "attention_qk_half_accumulation": false, |
| "bert_attention_plugin": "float16", |
| "context_fmha": false, |
| "context_fmha_fp32_acc": false, |
| "enable_xqa": false, |
| "gemm_plugin": "float16", |
| "gpt_attention_plugin": null, |
| "identity_plugin": null, |
| "layernorm_quantization_plugin": null, |
| "lookup_plugin": null, |
| "lora_plugin": null, |
| "mamba_conv1d_plugin": null, |
| "moe_plugin": null, |
| "multi_block_mode": false, |
| "multiple_profiles": false, |
| "nccl_plugin": null, |
| "paged_kv_cache": false, |
| "paged_state": false, |
| "quantize_per_token_plugin": false, |
| "quantize_tensor_plugin": false, |
| "remove_input_padding": false, |
| "rmsnorm_quantization_plugin": null, |
| "smooth_quant_gemm_plugin": null, |
| "streamingllm": false, |
| "tokens_per_block": 128, |
| "use_context_fmha_for_generation": false, |
| "use_custom_all_reduce": false, |
| "use_fp8_context_fmha": false, |
| "use_paged_context_fmha": false, |
| "weight_only_groupwise_quant_matmul_plugin": null, |
| "weight_only_quant_matmul_plugin": null |
| } |
| } |
|
|
