Upload folder using huggingface_hub

.gitattributes

@@ -41,3 +41,7 @@ h100-whisper-large-v3-trt-bs4/decoder.engine filter=lfs diff=lfs merge=lfs -text
 h100-whisper-large-v3-trt-bs4/encoder.engine filter=lfs diff=lfs merge=lfs -text
 h100-whisper-large-v3-trt-bs8/decoder.engine filter=lfs diff=lfs merge=lfs -text
 h100-whisper-large-v3-trt-bs8/encoder.engine filter=lfs diff=lfs merge=lfs -text
+h100-EraX-WoW-Turbo-V1.1-trt-bs4/decoder.engine filter=lfs diff=lfs merge=lfs -text
+h100-EraX-WoW-Turbo-V1.1-trt-bs4/encoder.engine filter=lfs diff=lfs merge=lfs -text
+h100-EraX-WoW-Turbo-V1.1-trt-bs8/decoder.engine filter=lfs diff=lfs merge=lfs -text
+h100-EraX-WoW-Turbo-V1.1-trt-bs8/encoder.engine filter=lfs diff=lfs merge=lfs -text

EraX-openai-format/pt_ckpt.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:09187401bca4da11b3c7d6f89e626ae424c1fb6f04f5cbb5e8ccf3c9b25fd063
+size 3235693084

h100-EraX-WoW-Turbo-V1.1-trt-bs4/decoder.engine

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f715de534831addd52ec4a18800f5a6f320cfec7d24323a34dda3edda2a9ec4d
+size 477856332

h100-EraX-WoW-Turbo-V1.1-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": 4,
+    "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
+  }
+}

h100-EraX-WoW-Turbo-V1.1-trt-bs4/encoder.engine

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9efbfeecdaaacb2be84f89b4f5d8fa0591f9214f01a5807778776ff2b08f58ac
+size 1278975860

h100-EraX-WoW-Turbo-V1.1-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
+  }
+}

h100-EraX-WoW-Turbo-V1.1-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": [9, 0], "cuda_device_name": "NVIDIA H100 80GB HBM3", "model_path": "models/EraX-openai/pt_ckpt.pt", "output_dir": "models/a5e58a1624ca0cf89ac71f4655448c0e"}

h100-EraX-WoW-Turbo-V1.1-trt-bs8/decoder.engine

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eabdb7abe1a69fd37798f350b8b42fc299eb67d269748e7bd81074dcbc31698d
+size 477855836

h100-EraX-WoW-Turbo-V1.1-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": 4,
+    "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
+  }
+}

h100-EraX-WoW-Turbo-V1.1-trt-bs8/encoder.engine

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:54b2a3b23791ab7592376aac3c8a5dff4004ea9cc07db682742bac5d88a183b5
+size 1278879492

h100-EraX-WoW-Turbo-V1.1-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
+  }
+}

h100-EraX-WoW-Turbo-V1.1-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": [9, 0], "cuda_device_name": "NVIDIA H100 80GB HBM3", "model_path": "models/EraX-openai/pt_ckpt.pt", "output_dir": "models/a1b4210556972e0b09557bee5c74b51e"}