Upload folder using huggingface_hub
Browse files- .gitattributes +2 -0
- decoder_config.json +57 -0
- encoder_config.json +50 -0
- whisper_decoder_float16_tp1_rank0.engine +3 -0
- whisper_encoder_float16_tp1_rank0.engine +3 -0
.gitattributes
CHANGED
@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
36 |
+
whisper_decoder_float16_tp1_rank0.engine filter=lfs diff=lfs merge=lfs -text
|
37 |
+
whisper_encoder_float16_tp1_rank0.engine filter=lfs diff=lfs merge=lfs -text
|
decoder_config.json
ADDED
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"builder_config": {
|
3 |
+
"apply_query_key_layer_scaling": false,
|
4 |
+
"cross_attention": true,
|
5 |
+
"has_position_embedding": true,
|
6 |
+
"has_token_type_embedding": false,
|
7 |
+
"hidden_act": "gelu",
|
8 |
+
"hidden_size": 1280,
|
9 |
+
"int8": true,
|
10 |
+
"max_batch_size": 8,
|
11 |
+
"max_beam_width": 2,
|
12 |
+
"max_input_len": 224,
|
13 |
+
"max_output_len": 512,
|
14 |
+
"max_position_embeddings": 448,
|
15 |
+
"name": "whisper_decoder",
|
16 |
+
"num_heads": 20,
|
17 |
+
"num_layers": 32,
|
18 |
+
"precision": "float16",
|
19 |
+
"strongly_typed": false,
|
20 |
+
"tensor_parallel": 1,
|
21 |
+
"use_refit": false,
|
22 |
+
"vocab_size": 51866
|
23 |
+
},
|
24 |
+
"plugin_config": {
|
25 |
+
"attention_qk_half_accumulation": false,
|
26 |
+
"bert_attention_plugin": null,
|
27 |
+
"context_fmha": true,
|
28 |
+
"context_fmha_fp32_acc": false,
|
29 |
+
"enable_xqa": false,
|
30 |
+
"gemm_plugin": "float16",
|
31 |
+
"gpt_attention_plugin": "float16",
|
32 |
+
"identity_plugin": null,
|
33 |
+
"layernorm_quantization_plugin": null,
|
34 |
+
"lookup_plugin": null,
|
35 |
+
"lora_plugin": null,
|
36 |
+
"mamba_conv1d_plugin": null,
|
37 |
+
"moe_plugin": null,
|
38 |
+
"multi_block_mode": false,
|
39 |
+
"multiple_profiles": false,
|
40 |
+
"nccl_plugin": null,
|
41 |
+
"paged_kv_cache": false,
|
42 |
+
"paged_state": false,
|
43 |
+
"quantize_per_token_plugin": false,
|
44 |
+
"quantize_tensor_plugin": false,
|
45 |
+
"remove_input_padding": false,
|
46 |
+
"rmsnorm_quantization_plugin": null,
|
47 |
+
"smooth_quant_gemm_plugin": null,
|
48 |
+
"streamingllm": false,
|
49 |
+
"tokens_per_block": 128,
|
50 |
+
"use_context_fmha_for_generation": false,
|
51 |
+
"use_custom_all_reduce": false,
|
52 |
+
"use_fp8_context_fmha": false,
|
53 |
+
"use_paged_context_fmha": false,
|
54 |
+
"weight_only_groupwise_quant_matmul_plugin": null,
|
55 |
+
"weight_only_quant_matmul_plugin": "float16"
|
56 |
+
}
|
57 |
+
}
|
encoder_config.json
ADDED
@@ -0,0 +1,50 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"builder_config": {
|
3 |
+
"hidden_size": 1280,
|
4 |
+
"int8": true,
|
5 |
+
"max_batch_size": 8,
|
6 |
+
"max_beam_width": 2,
|
7 |
+
"n_mels": 128,
|
8 |
+
"name": "whisper_encoder",
|
9 |
+
"num_heads": 20,
|
10 |
+
"num_languages": 100,
|
11 |
+
"num_layers": 32,
|
12 |
+
"precision": "float16",
|
13 |
+
"strongly_typed": false,
|
14 |
+
"tensor_parallel": 1,
|
15 |
+
"use_refit": false
|
16 |
+
},
|
17 |
+
"plugin_config": {
|
18 |
+
"attention_qk_half_accumulation": false,
|
19 |
+
"bert_attention_plugin": "float16",
|
20 |
+
"context_fmha": true,
|
21 |
+
"context_fmha_fp32_acc": false,
|
22 |
+
"enable_xqa": false,
|
23 |
+
"gemm_plugin": "float16",
|
24 |
+
"gpt_attention_plugin": null,
|
25 |
+
"identity_plugin": null,
|
26 |
+
"layernorm_quantization_plugin": null,
|
27 |
+
"lookup_plugin": null,
|
28 |
+
"lora_plugin": null,
|
29 |
+
"mamba_conv1d_plugin": null,
|
30 |
+
"moe_plugin": null,
|
31 |
+
"multi_block_mode": false,
|
32 |
+
"multiple_profiles": false,
|
33 |
+
"nccl_plugin": null,
|
34 |
+
"paged_kv_cache": false,
|
35 |
+
"paged_state": false,
|
36 |
+
"quantize_per_token_plugin": false,
|
37 |
+
"quantize_tensor_plugin": false,
|
38 |
+
"remove_input_padding": false,
|
39 |
+
"rmsnorm_quantization_plugin": null,
|
40 |
+
"smooth_quant_gemm_plugin": null,
|
41 |
+
"streamingllm": false,
|
42 |
+
"tokens_per_block": 128,
|
43 |
+
"use_context_fmha_for_generation": false,
|
44 |
+
"use_custom_all_reduce": false,
|
45 |
+
"use_fp8_context_fmha": false,
|
46 |
+
"use_paged_context_fmha": false,
|
47 |
+
"weight_only_groupwise_quant_matmul_plugin": null,
|
48 |
+
"weight_only_quant_matmul_plugin": "float16"
|
49 |
+
}
|
50 |
+
}
|
whisper_decoder_float16_tp1_rank0.engine
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:339e34b13e2563303125971953b06469ee3849a27730122ed1d97cca0376a5ed
|
3 |
+
size 1117804076
|
whisper_encoder_float16_tp1_rank0.engine
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a46f44cdaba8217b92e26b4eff1442817a14fa1d264c028fa51a69c906515d5b
|
3 |
+
size 650450564
|