imnotednamode commited on
Commit
d534305
1 Parent(s): 31b87c3
Files changed (2) hide show
  1. config.json +38 -0
  2. diffusion_pytorch_model.safetensors +3 -0
config.json ADDED
@@ -0,0 +1,38 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_class_name": "MochiTransformer3DModel",
3
+ "_diffusers_version": "0.32.0.dev0",
4
+ "_name_or_path": "mochi-1-diffusers",
5
+ "activation_fn": "swiglu",
6
+ "attention_head_dim": 128,
7
+ "in_channels": 12,
8
+ "max_sequence_length": 256,
9
+ "num_attention_heads": 24,
10
+ "num_layers": 48,
11
+ "out_channels": null,
12
+ "patch_size": 2,
13
+ "pooled_projection_dim": 1536,
14
+ "qk_norm": "rms_norm",
15
+ "quantization_config": {
16
+ "_load_in_4bit": true,
17
+ "_load_in_8bit": false,
18
+ "bnb_4bit_compute_dtype": "bfloat16",
19
+ "bnb_4bit_quant_storage": "uint8",
20
+ "bnb_4bit_quant_type": "nf4",
21
+ "bnb_4bit_use_double_quant": false,
22
+ "llm_int8_enable_fp32_cpu_offload": false,
23
+ "llm_int8_has_fp16_weight": false,
24
+ "llm_int8_skip_modules": [
25
+ "final_layer",
26
+ "x_embedder.proj",
27
+ "t_embedder",
28
+ "pos_frequencies",
29
+ "t5"
30
+ ],
31
+ "llm_int8_threshold": 6.0,
32
+ "load_in_4bit": true,
33
+ "load_in_8bit": false,
34
+ "quant_method": "bitsandbytes"
35
+ },
36
+ "text_embed_dim": 4096,
37
+ "time_embed_dim": 256
38
+ }
diffusion_pytorch_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2ea0ec312a48a6e29d8c94c6c85ed3c0e32363dbb3610558125d48ce375ea26d
3
+ size 5642936471