Hugging Face's logo

yuekai
/
whisper_turbo_and_distill_tllm_checkpoint

Model card Files Community
yuekai commited on
Commit
9500f83
•
1 Parent(s): 53c2b63

Upload folder using huggingface_hub

Browse files
Files changed (8) hide show
  1. distill_large_v3_tllm_checkpoint_encoder_pos_emb_true/decoder/config.json +38 -0
  2. distill_large_v3_tllm_checkpoint_encoder_pos_emb_true/decoder/rank0.safetensors +3 -0
  3. distill_large_v3_tllm_checkpoint_encoder_pos_emb_true/encoder/config.json +16 -0
  4. distill_large_v3_tllm_checkpoint_encoder_pos_emb_true/encoder/rank0.safetensors +3 -0
  5. tllm_checkpoint_turbo/decoder/config.json +38 -0
  6. tllm_checkpoint_turbo/decoder/rank0.safetensors +3 -0
  7. tllm_checkpoint_turbo/encoder/config.json +16 -0
  8. tllm_checkpoint_turbo/encoder/rank0.safetensors +3 -0
distill_large_v3_tllm_checkpoint_encoder_pos_emb_true/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_large_v3_tllm_checkpoint_encoder_pos_emb_true/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_large_v3_tllm_checkpoint_encoder_pos_emb_true/encoder/config.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architecture": "WhisperEncoder",
3
+ "dtype": "float16",
4
+ "num_hidden_layers": 32,
5
+ "num_attention_heads": 20,
6
+ "hidden_size": 1280,
7
+ "max_position_embeddings": 1500,
8
+ "has_position_embedding": true,
9
+ "n_mels": 128,
10
+ "vocab_size": 51866,
11
+ "hidden_act": "gelu",
12
+ "num_languages": 100,
13
+ "quantization": {
14
+ "quant_algo": null
15
+ }
16
+ }
distill_large_v3_tllm_checkpoint_encoder_pos_emb_true/encoder/rank0.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bbbf025e3626bd5e72fdbd8386f16518bf2684aec2f9a87d6819d523c28de02f
3
+ size 1277902304
tllm_checkpoint_turbo/decoder/config.json ADDED
@@ -0,0 +1,38 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architecture": "DecoderModel",
3
+ "dtype": "float16",
4
+ "logits_dtype": "float16",
5
+ "num_hidden_layers": 4,
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
+ }
tllm_checkpoint_turbo/decoder/rank0.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6c93c16de2a985952babbbff292f4aac98189046249c9e096c97c3204f28ade4
3
+ size 476624224
tllm_checkpoint_turbo/encoder/config.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architecture": "WhisperEncoder",
3
+ "dtype": "float16",
4
+ "num_hidden_layers": 32,
5
+ "num_attention_heads": 20,
6
+ "hidden_size": 1280,
7
+ "max_position_embeddings": 1500,
8
+ "has_position_embedding": true,
9
+ "n_mels": 128,
10
+ "vocab_size": 51866,
11
+ "hidden_act": "gelu",
12
+ "num_languages": 100,
13
+ "quantization": {
14
+ "quant_algo": null
15
+ }
16
+ }
tllm_checkpoint_turbo/encoder/rank0.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a40e7340cba791bc022f4eee07a53acb38d5196838b6dd667d2027eceee98fd3
3
+ size 1277902304