Upload BNB8 quantized FlashTalk DiT (bitsandbytes)

bnb8/config.json

@@ -0,0 +1,38 @@
+{
+  "_class_name": "WanModel",
+  "_diffusers_version": "0.37.0.dev0",
+  "_name_or_path": "/models/SoulX-FlashTalk-14B",
+  "audio_window": 5,
+  "context_tokens": 32,
+  "dim": 5120,
+  "eps": 1e-06,
+  "ffn_dim": 13824,
+  "freq_dim": 256,
+  "in_dim": 36,
+  "intermediate_dim": 512,
+  "model_type": "i2v",
+  "norm_input_visual": true,
+  "norm_output_audio": true,
+  "num_heads": 40,
+  "num_layers": 40,
+  "out_dim": 16,
+  "output_dim": 768,
+  "quantization_config": {
+    "_load_in_4bit": false,
+    "_load_in_8bit": true,
+    "bnb_4bit_compute_dtype": "float32",
+    "bnb_4bit_quant_storage": "uint8",
+    "bnb_4bit_quant_type": "fp4",
+    "bnb_4bit_use_double_quant": false,
+    "llm_int8_enable_fp32_cpu_offload": false,
+    "llm_int8_has_fp16_weight": false,
+    "llm_int8_skip_modules": null,
+    "llm_int8_threshold": 6.0,
+    "load_in_4bit": false,
+    "load_in_8bit": true,
+    "quant_method": "bitsandbytes"
+  },
+  "text_len": 512,
+  "vae_scale": 4,
+  "weight_init": false
+}

bnb8/diffusion_pytorch_model-00001-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0e520d37c81f14d20eae397750faf7b284541c0102d78e10f0c4181e8fd5620e
+size 9992537856

bnb8/diffusion_pytorch_model-00002-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a2a79269192d7a453fa1a41efc2fa8a7b2be8138c5fd736929053f852684fcd5
+size 8912179428

bnb8/flashtalk_bnb8_meta.json

@@ -0,0 +1,5 @@
+{
+  "config": "BitsAndBytesConfig",
+  "variant": "bnb8",
+  "version": 1
+}