Engine build from baseten-whisper@b0e7e62615487eb9b5be6e0c58c18e5eaa2cf66f (trtllm 1.3.0rc20)

.gitattributes

@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+decoder/rank0.engine filter=lfs diff=lfs merge=lfs -text
+encoder/rank0.engine filter=lfs diff=lfs merge=lfs -text

decoder/config.json

@@ -0,0 +1,161 @@
+{
+    "version": "1.3.0rc20",
+    "pretrained_config": {
+        "architecture": "DecoderModel",
+        "dtype": "float16",
+        "vocab_size": 51865,
+        "hidden_size": 1280,
+        "num_hidden_layers": 32,
+        "num_attention_heads": 20,
+        "hidden_act": "gelu",
+        "logits_dtype": "float16",
+        "norm_epsilon": 1e-05,
+        "runtime_defaults": null,
+        "position_embedding_type": "learned_absolute",
+        "num_key_value_heads": 20,
+        "intermediate_size": 5120,
+        "max_position_embeddings": 448,
+        "mapping": {
+            "world_size": 1,
+            "gpus_per_node": 8,
+            "cp_size": 1,
+            "tp_size": 1,
+            "pp_size": 1,
+            "moe_tp_size": 1,
+            "moe_cluster_size": 1,
+            "moe_ep_size": 1,
+            "attn_tp_size": 1,
+            "attn_cp_size": 1,
+            "cp_config": {},
+            "enable_attention_dp": false,
+            "enable_lm_head_tp_in_adp": false,
+            "dwdp_size": 0,
+            "dwdp_rank": 0
+        },
+        "quantization": {
+            "quant_algo": "W8A16",
+            "kv_cache_quant_algo": null,
+            "group_size": 128,
+            "smoothquant_val": 0.5,
+            "clamp_val": null,
+            "use_meta_recipe": false,
+            "has_zero_point": false,
+            "pre_quant_scale": false,
+            "exclude_modules": null,
+            "mamba_ssm_cache_dtype": null,
+            "mamba_ssm_stochastic_rounding": false,
+            "mamba_ssm_philox_rounds": 10
+        },
+        "use_parallel_embedding": false,
+        "embedding_sharding_dim": 0,
+        "head_size": 64,
+        "qk_layernorm": false,
+        "rotary_embedding_dim": 64,
+        "use_prompt_tuning": false,
+        "has_position_embedding": true,
+        "layernorm_type": 0,
+        "has_attention_qkvo_bias": true,
+        "has_mlp_bias": true,
+        "has_model_final_layernorm": true,
+        "has_embedding_layernorm": false,
+        "has_embedding_scale": false,
+        "ffn_hidden_size": 5120,
+        "q_scaling": 1.0,
+        "layernorm_position": 0,
+        "relative_attention": false,
+        "max_distance": 0,
+        "num_buckets": 0,
+        "model_type": "whisper",
+        "rescale_before_lm_head": false,
+        "encoder_hidden_size": 1280,
+        "encoder_num_heads": 20,
+        "encoder_head_size": null,
+        "skip_cross_kv": false,
+        "type_vocab_size": null,
+        "encoder_num_kv_heads": null,
+        "mlp_type": 0,
+        "residual_scaling": 1.0,
+        "has_lm_head_bias": false
+    },
+    "build_config": {
+        "max_input_len": 100,
+        "max_seq_len": 400,
+        "opt_batch_size": 8,
+        "max_batch_size": 14,
+        "max_beam_width": 5,
+        "max_num_tokens": 5600,
+        "opt_num_tokens": 70,
+        "max_prompt_embedding_table_size": 0,
+        "kv_cache_type": "paged",
+        "gather_context_logits": false,
+        "gather_generation_logits": true,
+        "strongly_typed": true,
+        "force_num_profiles": null,
+        "profiling_verbosity": "layer_names_only",
+        "enable_debug_output": false,
+        "max_draft_len": 0,
+        "speculative_decoding_mode": 1,
+        "use_refit": false,
+        "input_timing_cache": null,
+        "output_timing_cache": "model.cache",
+        "lora_config": {
+            "lora_dir": [],
+            "lora_ckpt_source": "hf",
+            "max_lora_rank": 64,
+            "lora_target_modules": [],
+            "trtllm_modules_to_hf_modules": {},
+            "max_loras": null,
+            "max_cpu_loras": null,
+            "swap_gate_up_proj_lora_b_weight": true
+        },
+        "weight_sparsity": false,
+        "weight_streaming": false,
+        "plugin_config": {
+            "dtype": "float16",
+            "bert_attention_plugin": "float16",
+            "gpt_attention_plugin": "float16",
+            "gemm_plugin": "float16",
+            "gemm_swiglu_plugin": null,
+            "fp8_rowwise_gemm_plugin": null,
+            "qserve_gemm_plugin": null,
+            "identity_plugin": null,
+            "nccl_plugin": null,
+            "lora_plugin": null,
+            "dora_plugin": false,
+            "weight_only_groupwise_quant_matmul_plugin": null,
+            "weight_only_quant_matmul_plugin": "float16",
+            "smooth_quant_plugins": true,
+            "smooth_quant_gemm_plugin": null,
+            "layernorm_quantization_plugin": null,
+            "rmsnorm_quantization_plugin": null,
+            "quantize_per_token_plugin": false,
+            "quantize_tensor_plugin": false,
+            "moe_plugin": null,
+            "mamba_conv1d_plugin": "auto",
+            "low_latency_gemm_plugin": null,
+            "low_latency_gemm_swiglu_plugin": null,
+            "gemm_allreduce_plugin": null,
+            "context_fmha": true,
+            "bert_context_fmha_fp32_acc": false,
+            "paged_kv_cache": true,
+            "remove_input_padding": true,
+            "norm_quant_fusion": false,
+            "reduce_fusion": false,
+            "user_buffer": false,
+            "tokens_per_block": 32,
+            "use_paged_context_fmha": true,
+            "use_fp8_context_fmha": false,
+            "fuse_fp4_quant": false,
+            "multiple_profiles": false,
+            "paged_state": false,
+            "streamingllm": false,
+            "manage_weights": false,
+            "use_fused_mlp": true,
+            "pp_reduce_scatter": false
+        },
+        "use_strip_plan": false,
+        "max_encoder_input_len": 3000,
+        "monitor_memory": false,
+        "use_mrope": false
+    }
+}

decoder/rank0.engine

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:059c58a33ad25b8e61c42d70a92033b808ac166bc3a18b13401a440c0fdc3648
+size 1119764772

encoder/config.json

@@ -0,0 +1,139 @@
+{
+    "version": "1.3.0rc20",
+    "pretrained_config": {
+        "architecture": "WhisperEncoder",
+        "dtype": "float16",
+        "vocab_size": 51865,
+        "hidden_size": 1280,
+        "num_hidden_layers": 32,
+        "num_attention_heads": 20,
+        "hidden_act": "gelu",
+        "logits_dtype": "float32",
+        "norm_epsilon": 1e-05,
+        "runtime_defaults": null,
+        "position_embedding_type": "learned_absolute",
+        "num_key_value_heads": 20,
+        "intermediate_size": 5120,
+        "max_position_embeddings": 1500,
+        "mapping": {
+            "world_size": 1,
+            "gpus_per_node": 8,
+            "cp_size": 1,
+            "tp_size": 1,
+            "pp_size": 1,
+            "moe_tp_size": 1,
+            "moe_cluster_size": 1,
+            "moe_ep_size": 1,
+            "attn_tp_size": 1,
+            "attn_cp_size": 1,
+            "cp_config": {},
+            "enable_attention_dp": false,
+            "enable_lm_head_tp_in_adp": false,
+            "dwdp_size": 0,
+            "dwdp_rank": 0
+        },
+        "quantization": {
+            "quant_algo": "W8A16",
+            "kv_cache_quant_algo": null,
+            "group_size": 128,
+            "smoothquant_val": 0.5,
+            "clamp_val": null,
+            "use_meta_recipe": false,
+            "has_zero_point": false,
+            "pre_quant_scale": false,
+            "exclude_modules": null,
+            "mamba_ssm_cache_dtype": null,
+            "mamba_ssm_stochastic_rounding": false,
+            "mamba_ssm_philox_rounds": 10
+        },
+        "use_parallel_embedding": false,
+        "embedding_sharding_dim": 0,
+        "head_size": 64,
+        "qk_layernorm": false,
+        "rotary_embedding_dim": 64,
+        "has_position_embedding": true,
+        "n_mels": 80,
+        "num_languages": 99
+    },
+    "build_config": {
+        "max_input_len": 3000,
+        "max_seq_len": 3000,
+        "opt_batch_size": 8,
+        "max_batch_size": 14,
+        "max_beam_width": 1,
+        "max_num_tokens": 8192,
+        "opt_num_tokens": 14,
+        "max_prompt_embedding_table_size": 0,
+        "kv_cache_type": "paged",
+        "gather_context_logits": false,
+        "gather_generation_logits": false,
+        "strongly_typed": true,
+        "force_num_profiles": null,
+        "profiling_verbosity": "layer_names_only",
+        "enable_debug_output": false,
+        "max_draft_len": 0,
+        "speculative_decoding_mode": 1,
+        "use_refit": false,
+        "input_timing_cache": null,
+        "output_timing_cache": "model.cache",
+        "lora_config": {
+            "lora_dir": [],
+            "lora_ckpt_source": "hf",
+            "max_lora_rank": 64,
+            "lora_target_modules": [],
+            "trtllm_modules_to_hf_modules": {},
+            "max_loras": null,
+            "max_cpu_loras": null,
+            "swap_gate_up_proj_lora_b_weight": true
+        },
+        "weight_sparsity": false,
+        "weight_streaming": false,
+        "plugin_config": {
+            "dtype": "float16",
+            "bert_attention_plugin": "float16",
+            "gpt_attention_plugin": "auto",
+            "gemm_plugin": null,
+            "gemm_swiglu_plugin": null,
+            "fp8_rowwise_gemm_plugin": null,
+            "qserve_gemm_plugin": null,
+            "identity_plugin": null,
+            "nccl_plugin": null,
+            "lora_plugin": null,
+            "dora_plugin": false,
+            "weight_only_groupwise_quant_matmul_plugin": null,
+            "weight_only_quant_matmul_plugin": "float16",
+            "smooth_quant_plugins": true,
+            "smooth_quant_gemm_plugin": null,
+            "layernorm_quantization_plugin": null,
+            "rmsnorm_quantization_plugin": null,
+            "quantize_per_token_plugin": false,
+            "quantize_tensor_plugin": false,
+            "moe_plugin": null,
+            "mamba_conv1d_plugin": "auto",
+            "low_latency_gemm_plugin": null,
+            "low_latency_gemm_swiglu_plugin": null,
+            "gemm_allreduce_plugin": null,
+            "context_fmha": true,
+            "bert_context_fmha_fp32_acc": false,
+            "paged_kv_cache": true,
+            "remove_input_padding": true,
+            "norm_quant_fusion": false,
+            "reduce_fusion": false,
+            "user_buffer": false,
+            "tokens_per_block": 32,
+            "use_paged_context_fmha": true,
+            "use_fp8_context_fmha": false,
+            "fuse_fp4_quant": false,
+            "multiple_profiles": false,
+            "paged_state": false,
+            "streamingllm": false,
+            "manage_weights": false,
+            "use_fused_mlp": true,
+            "pp_reduce_scatter": false
+        },
+        "use_strip_plan": false,
+        "max_encoder_input_len": 1024,
+        "monitor_memory": false,
+        "use_mrope": false
+    }
+}

encoder/rank0.engine

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eca45cdd5afae848e0c38caf0a0aeacfcbd316ae63214c8ea07ff2a4f8bba29d
+size 656750748