Hugging Face's logo

yuekai
/
distill_whisper_large_v3_trtllm_triton

Model card Files Community
yuekai commited on
Commit
c398568
•
1 Parent(s): f605912

Upload folder using huggingface_hub

Browse files
Files changed (8) hide show
  1. distill_tllm_checkpoint/decoder/config.json +38 -0
  2. distill_tllm_checkpoint/decoder/rank0.safetensors +3 -0
  3. distill_tllm_checkpoint/encoder/config.json +15 -0
  4. distill_tllm_checkpoint/encoder/rank0.safetensors +3 -0
  5. distill_tllm_checkpoint_int8/decoder/config.json +38 -0
  6. distill_tllm_checkpoint_int8/decoder/rank0.safetensors +3 -0
  7. distill_tllm_checkpoint_int8/encoder/config.json +15 -0
  8. distill_tllm_checkpoint_int8/encoder/rank0.safetensors +3 -0
distill_tllm_checkpoint/decoder/config.json ADDED
@@ -0,0 +1,38 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architecture": "DecoderModel",
3
+ "dtype": "float16",
4
+ "logits_dtype": "float16",
5
+ "num_hidden_layers": 2,
6
+ "num_attention_heads": 20,
7
+ "hidden_size": 1280,
8
+ "norm_epsilon": 1e-05,
9
+ "vocab_size": 51866,
10
+ "hidden_act": "gelu",
11
+ "use_parallel_embedding": false,
12
+ "embedding_sharding_dim": 0,
13
+ "max_position_embeddings": 448,
14
+ "use_prompt_tuning": false,
15
+ "head_size": 64,
16
+ "has_position_embedding": true,
17
+ "layernorm_type": 0,
18
+ "has_attention_qkvo_bias": true,
19
+ "has_mlp_bias": true,
20
+ "has_model_final_layernorm": true,
21
+ "has_embedding_layernorm": false,
22
+ "has_embedding_scale": false,
23
+ "ffn_hidden_size": 5120,
24
+ "q_scaling": 1.0,
25
+ "layernorm_position": 0,
26
+ "relative_attention": false,
27
+ "max_distance": 0,
28
+ "num_buckets": 0,
29
+ "model_type": "whisper",
30
+ "rescale_before_lm_head": false,
31
+ "encoder_hidden_size": 1280,
32
+ "encoder_num_heads": 20,
33
+ "encoder_head_size": null,
34
+ "skip_cross_qkv": false,
35
+ "quantization": {
36
+ "quant_algo": null
37
+ }
38
+ }
distill_tllm_checkpoint/decoder/rank0.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1a2e1b9d985e4764effd3367fb7994c5b5cc4ac1dedd63f3c829d05f30113118
3
+ size 371665280
distill_tllm_checkpoint/encoder/config.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architecture": "WhisperEncoder",
3
+ "dtype": "float16",
4
+ "num_hidden_layers": 32,
5
+ "num_attention_heads": 20,
6
+ "hidden_size": 1280,
7
+ "n_mels": 128,
8
+ "n_audio_ctx": 1500,
9
+ "vocab_size": 51866,
10
+ "hidden_act": "gelu",
11
+ "num_languages": 100,
12
+ "quantization": {
13
+ "quant_algo": null
14
+ }
15
+ }
distill_tllm_checkpoint/encoder/rank0.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:68cea668723d154ea21c04d6afe844714473672575894b496b1f604e892ebddc
3
+ size 1277902296
distill_tllm_checkpoint_int8/decoder/config.json ADDED
@@ -0,0 +1,38 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architecture": "DecoderModel",
3
+ "dtype": "float16",
4
+ "logits_dtype": "float16",
5
+ "num_hidden_layers": 2,
6
+ "num_attention_heads": 20,
7
+ "hidden_size": 1280,
8
+ "norm_epsilon": 1e-05,
9
+ "vocab_size": 51866,
10
+ "hidden_act": "gelu",
11
+ "use_parallel_embedding": false,
12
+ "embedding_sharding_dim": 0,
13
+ "max_position_embeddings": 448,
14
+ "use_prompt_tuning": false,
15
+ "head_size": 64,
16
+ "has_position_embedding": true,
17
+ "layernorm_type": 0,
18
+ "has_attention_qkvo_bias": true,
19
+ "has_mlp_bias": true,
20
+ "has_model_final_layernorm": true,
21
+ "has_embedding_layernorm": false,
22
+ "has_embedding_scale": false,
23
+ "ffn_hidden_size": 5120,
24
+ "q_scaling": 1.0,
25
+ "layernorm_position": 0,
26
+ "relative_attention": false,
27
+ "max_distance": 0,
28
+ "num_buckets": 0,
29
+ "model_type": "whisper",
30
+ "rescale_before_lm_head": false,
31
+ "encoder_hidden_size": 1280,
32
+ "encoder_num_heads": 20,
33
+ "encoder_head_size": null,
34
+ "skip_cross_qkv": false,
35
+ "quantization": {
36
+ "quant_algo": "W8A16"
37
+ }
38
+ }
distill_tllm_checkpoint_int8/decoder/rank0.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d244e1a0ef5ac353b91cb44f12e962bad8bbca2032e89c3e32d4359bc777a64c
3
+ size 319304296
distill_tllm_checkpoint_int8/encoder/config.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architecture": "WhisperEncoder",
3
+ "dtype": "float16",
4
+ "num_hidden_layers": 32,
5
+ "num_attention_heads": 20,
6
+ "hidden_size": 1280,
7
+ "n_mels": 128,
8
+ "n_audio_ctx": 1500,
9
+ "vocab_size": 51866,
10
+ "hidden_act": "gelu",
11
+ "num_languages": 100,
12
+ "quantization": {
13
+ "quant_algo": "W8A16"
14
+ }
15
+ }
distill_tllm_checkpoint_int8/encoder/rank0.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3680afaeeb22b9b2ed68895dfcc8b8e7e09d8e443789d7c227ada92ee2b36860
3
+ size 649507920