Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,7 @@ 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
+4090-whisper-large-v3-trt-bs4/decoder.engine filter=lfs diff=lfs merge=lfs -text
+4090-whisper-large-v3-trt-bs4/encoder.engine filter=lfs diff=lfs merge=lfs -text
+4090-whisper-large-v3-trt-bs8/decoder.engine filter=lfs diff=lfs merge=lfs -text
+4090-whisper-large-v3-trt-bs8/encoder.engine filter=lfs diff=lfs merge=lfs -text

4090-whisper-large-v3-trt-bs4/decoder.engine

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:00eaed8abe5a70fadc4046cdbc01d8b0964f96d723a1bcb4687ca46d1cb77e31
+size 1954679476

4090-whisper-large-v3-trt-bs4/decoder_config.json

@@ -0,0 +1,51 @@
+{
+  "builder_config": {
+    "apply_query_key_layer_scaling": false,
+    "cross_attention": true,
+    "has_position_embedding": true,
+    "has_token_type_embedding": false,
+    "hidden_act": "gelu",
+    "hidden_size": 1280,
+    "int8": false,
+    "max_batch_size": 4,
+    "max_input_len": 14,
+    "max_output_len": 448,
+    "max_position_embeddings": 448,
+    "name": "decoder",
+    "num_heads": 20,
+    "num_layers": 32,
+    "precision": "float16",
+    "tensor_parallel": 1,
+    "use_refit": false,
+    "vocab_size": 51866
+  },
+  "plugin_config": {
+    "attention_qk_half_accumulation": false,
+    "bert_attention_plugin": null,
+    "context_fmha_type": 0,
+    "enable_xqa": false,
+    "gemm_plugin": "float16",
+    "gpt_attention_plugin": "float16",
+    "identity_plugin": null,
+    "layernorm_plugin": null,
+    "layernorm_quantization_plugin": null,
+    "lookup_plugin": null,
+    "lora_plugin": null,
+    "multi_block_mode": false,
+    "nccl_plugin": null,
+    "paged_kv_cache": false,
+    "quantize_per_token_plugin": false,
+    "quantize_tensor_plugin": false,
+    "remove_input_padding": false,
+    "rmsnorm_plugin": null,
+    "rmsnorm_quantization_plugin": null,
+    "selective_scan_plugin": false,
+    "smooth_quant_gemm_plugin": null,
+    "tokens_per_block": 0,
+    "use_context_fmha_for_generation": false,
+    "use_custom_all_reduce": false,
+    "use_paged_context_fmha": false,
+    "weight_only_groupwise_quant_matmul_plugin": null,
+    "weight_only_quant_matmul_plugin": null
+  }
+}

4090-whisper-large-v3-trt-bs4/encoder.engine

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:29e48a254ec42ce4d33d6b091608b3970b29c4058bf68a21838f8c439952cdb4
+size 1278471588

4090-whisper-large-v3-trt-bs4/encoder_config.json

@@ -0,0 +1,44 @@
+{
+  "builder_config": {
+    "hidden_size": 1280,
+    "int8": false,
+    "max_batch_size": 4,
+    "n_mels": 128,
+    "name": "encoder",
+    "num_heads": 20,
+    "num_languages": 100,
+    "num_layers": 32,
+    "precision": "float16",
+    "tensor_parallel": 1,
+    "use_refit": false
+  },
+  "plugin_config": {
+    "attention_qk_half_accumulation": false,
+    "bert_attention_plugin": "float16",
+    "context_fmha_type": 0,
+    "enable_xqa": false,
+    "gemm_plugin": "float16",
+    "gpt_attention_plugin": null,
+    "identity_plugin": null,
+    "layernorm_plugin": null,
+    "layernorm_quantization_plugin": null,
+    "lookup_plugin": null,
+    "lora_plugin": null,
+    "multi_block_mode": false,
+    "nccl_plugin": null,
+    "paged_kv_cache": false,
+    "quantize_per_token_plugin": false,
+    "quantize_tensor_plugin": false,
+    "remove_input_padding": false,
+    "rmsnorm_plugin": null,
+    "rmsnorm_quantization_plugin": null,
+    "selective_scan_plugin": false,
+    "smooth_quant_gemm_plugin": null,
+    "tokens_per_block": 0,
+    "use_context_fmha_for_generation": false,
+    "use_custom_all_reduce": false,
+    "use_paged_context_fmha": false,
+    "weight_only_groupwise_quant_matmul_plugin": null,
+    "weight_only_quant_matmul_plugin": null
+  }
+}

4090-whisper-large-v3-trt-bs4/trt_build_args.json

@@ -0,0 +1 @@
+{"max_batch_size": 4, "max_beam_width": 1, "max_input_len": 14, "max_output_len": 448, "world_size": 1, "dtype": "float16", "quantize_dir": "quantize/1-gpu", "use_gpt_attention_plugin": "float16", "use_bert_attention_plugin": "float16", "use_context_fmha_enc": false, "use_context_fmha_dec": false, "use_gemm_plugin": "float16", "use_layernorm_plugin": false, "remove_input_padding": false, "use_weight_only_enc": false, "use_weight_only_dec": false, "weight_only_precision": "int8", "int8_kv_cache": false, "debug_mode": false, "cuda_compute_capability": [8, 9], "cuda_device_name": "NVIDIA GeForce RTX 4090", "model_path": "models/large-v3/pt_ckpt.pt", "output_dir": "models/266044c7309ecb70c167ed90aee5bbf1"}

4090-whisper-large-v3-trt-bs8/decoder.engine

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5c275ffc8a50e85830d771da5b66d5eb6e299bad3a439a3b9338b7f4f4bb01c4
+size 1954647324

4090-whisper-large-v3-trt-bs8/decoder_config.json

@@ -0,0 +1,51 @@
+{
+  "builder_config": {
+    "apply_query_key_layer_scaling": false,
+    "cross_attention": true,
+    "has_position_embedding": true,
+    "has_token_type_embedding": false,
+    "hidden_act": "gelu",
+    "hidden_size": 1280,
+    "int8": false,
+    "max_batch_size": 8,
+    "max_input_len": 14,
+    "max_output_len": 448,
+    "max_position_embeddings": 448,
+    "name": "decoder",
+    "num_heads": 20,
+    "num_layers": 32,
+    "precision": "float16",
+    "tensor_parallel": 1,
+    "use_refit": false,
+    "vocab_size": 51866
+  },
+  "plugin_config": {
+    "attention_qk_half_accumulation": false,
+    "bert_attention_plugin": null,
+    "context_fmha_type": 0,
+    "enable_xqa": false,
+    "gemm_plugin": "float16",
+    "gpt_attention_plugin": "float16",
+    "identity_plugin": null,
+    "layernorm_plugin": null,
+    "layernorm_quantization_plugin": null,
+    "lookup_plugin": null,
+    "lora_plugin": null,
+    "multi_block_mode": false,
+    "nccl_plugin": null,
+    "paged_kv_cache": false,
+    "quantize_per_token_plugin": false,
+    "quantize_tensor_plugin": false,
+    "remove_input_padding": false,
+    "rmsnorm_plugin": null,
+    "rmsnorm_quantization_plugin": null,
+    "selective_scan_plugin": false,
+    "smooth_quant_gemm_plugin": null,
+    "tokens_per_block": 0,
+    "use_context_fmha_for_generation": false,
+    "use_custom_all_reduce": false,
+    "use_paged_context_fmha": false,
+    "weight_only_groupwise_quant_matmul_plugin": null,
+    "weight_only_quant_matmul_plugin": null
+  }
+}

4090-whisper-large-v3-trt-bs8/encoder.engine

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be6afdcdbab6b7380ae0bfb0da389b2ff86c4851083a425c08993532b38bb455
+size 1278764844

4090-whisper-large-v3-trt-bs8/encoder_config.json

@@ -0,0 +1,44 @@
+{
+  "builder_config": {
+    "hidden_size": 1280,
+    "int8": false,
+    "max_batch_size": 8,
+    "n_mels": 128,
+    "name": "encoder",
+    "num_heads": 20,
+    "num_languages": 100,
+    "num_layers": 32,
+    "precision": "float16",
+    "tensor_parallel": 1,
+    "use_refit": false
+  },
+  "plugin_config": {
+    "attention_qk_half_accumulation": false,
+    "bert_attention_plugin": "float16",
+    "context_fmha_type": 0,
+    "enable_xqa": false,
+    "gemm_plugin": "float16",
+    "gpt_attention_plugin": null,
+    "identity_plugin": null,
+    "layernorm_plugin": null,
+    "layernorm_quantization_plugin": null,
+    "lookup_plugin": null,
+    "lora_plugin": null,
+    "multi_block_mode": false,
+    "nccl_plugin": null,
+    "paged_kv_cache": false,
+    "quantize_per_token_plugin": false,
+    "quantize_tensor_plugin": false,
+    "remove_input_padding": false,
+    "rmsnorm_plugin": null,
+    "rmsnorm_quantization_plugin": null,
+    "selective_scan_plugin": false,
+    "smooth_quant_gemm_plugin": null,
+    "tokens_per_block": 0,
+    "use_context_fmha_for_generation": false,
+    "use_custom_all_reduce": false,
+    "use_paged_context_fmha": false,
+    "weight_only_groupwise_quant_matmul_plugin": null,
+    "weight_only_quant_matmul_plugin": null
+  }
+}

4090-whisper-large-v3-trt-bs8/trt_build_args.json

@@ -0,0 +1 @@
+{"max_batch_size": 8, "max_beam_width": 1, "max_input_len": 14, "max_output_len": 448, "world_size": 1, "dtype": "float16", "quantize_dir": "quantize/1-gpu", "use_gpt_attention_plugin": "float16", "use_bert_attention_plugin": "float16", "use_context_fmha_enc": false, "use_context_fmha_dec": false, "use_gemm_plugin": "float16", "use_layernorm_plugin": false, "remove_input_padding": false, "use_weight_only_enc": false, "use_weight_only_dec": false, "weight_only_precision": "int8", "int8_kv_cache": false, "debug_mode": false, "cuda_compute_capability": [8, 9], "cuda_device_name": "NVIDIA GeForce RTX 4090", "model_path": "models/large-v3/pt_ckpt.pt", "output_dir": "models/e19300457c5c10b1afb8832c3b56157f"}