API tools faq
paste
Advertisement
Guest User

Colab Framepack error

a guest
May 1st, 2025
63
0
Never
Add comment
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 17.69 KB | Source Code | 0 0
raw download clone embed print report
  1. Currently enabled native sdp backends: ['flash', 'math', 'mem_efficient', 'cudnn']
  2. Xformers is not installed!
  3. Flash Attn is not installed!
  4. Sage Attn is installed!
  5. Namespace(share=True, server='0.0.0.0', port=None, inbrowser=False)
  6. Free VRAM 14.6407470703125 GB
  7. High-VRAM Mode: False
  8. Downloading shards: 100% 4/4 [00:00<00:00, 13127.71it/s]
  9. Loading checkpoint shards: 100% 4/4 [00:01<00:00, 3.79it/s]
  10. Fetching 3 files: 100% 3/3 [00:00<00:00, 8045.34it/s]
  11. Loading checkpoint shards: 100% 3/3 [00:01<00:00, 1.65it/s]
  12. transformer.high_quality_fp32_output_for_inference = True
  13. * Running on local URL: http://
  14. * Running on public URL: https://
  15.  
  16. This share link expires in 72 hours. For free permanent hosting and GPU upgrades, run `gradio deploy` from the terminal in the working directory to deploy to Hugging Face Spaces (https://huggingface.co/spaces)
  17. Unloaded DynamicSwap_LlamaModel as complete.
  18. Unloaded CLIPTextModel as complete.
  19. Unloaded SiglipVisionModel as complete.
  20. Unloaded AutoencoderKLHunyuanVideo as complete.
  21. Unloaded DynamicSwap_HunyuanVideoTransformer3DModelPacked as complete.
  22. Loaded CLIPTextModel to cuda:0 as complete.
  23. Unloaded CLIPTextModel as complete.
  24. Loaded AutoencoderKLHunyuanVideo to cuda:0 as complete.
  25. Unloaded AutoencoderKLHunyuanVideo as complete.
  26. Loaded SiglipVisionModel to cuda:0 as complete.
  27. latent_padding_size = 27, is_last_section = False
  28. Unloaded SiglipVisionModel as complete.
  29. Moving DynamicSwap_HunyuanVideoTransformer3DModelPacked to cuda:0 with preserved memory: 6 GB
  30. 0% 0/25 [00:00<?, ?it/s]/usr/local/lib/python3.10/dist-packages/sageattention/attn_qk_int8_per_block.py:18:23: error: 'tt.fp_to_fp' op operand #0 must be floating-point or ranked tensor of floating-point values, but got 'tensor<128x128xi8, #ttg.dot_op<{opIdx = 0, parent = #ttg.blocked<{sizePerThread = [4, 4], threadsPerWarp = [2, 16], warpsPerCTA = [8, 1], order = [1, 0]}>}>>'
  31. qk = tl.dot(q, k).to(tl.float32) * q_scale * k_scale
  32. ^
  33. /usr/local/lib/python3.10/dist-packages/sageattention/attn_qk_int8_per_block.py:78:55: note: called from
  34. 4 - STAGE, offs_m, offs_n
  35. ^
  36. module {
  37. tt.func public @_attn_fwd(%arg0: !tt.ptr<i8> {tt.divisibility = 16 : i32}, %arg1: !tt.ptr<i8> {tt.divisibility = 16 : i32}, %arg2: !tt.ptr<f16> {tt.divisibility = 16 : i32}, %arg3: !tt.ptr<f32> {tt.divisibility = 16 : i32}, %arg4: !tt.ptr<f32> {tt.divisibility = 16 : i32}, %arg5: !tt.ptr<bf16> {tt.divisibility = 16 : i32}, %arg6: i32 {tt.divisibility = 16 : i32}, %arg7: i32 {tt.divisibility = 16 : i32}, %arg8: i32 {tt.divisibility = 16 : i32}, %arg9: i32 {tt.divisibility = 16 : i32}, %arg10: i32 {tt.divisibility = 16 : i32}, %arg11: i32 {tt.divisibility = 16 : i32}, %arg12: i32 {tt.divisibility = 16 : i32}, %arg13: i32 {tt.divisibility = 16 : i32}, %arg14: i32 {tt.divisibility = 16 : i32}, %arg15: i32 {tt.divisibility = 16 : i32}, %arg16: i32 {tt.divisibility = 16 : i32}, %arg17: i32 {tt.divisibility = 16 : i32}, %arg18: i32, %arg19: i32) attributes {noinline = false} {
  38. %cst = arith.constant dense<1.000000e+00> : tensor<128xf32>
  39. %cst_0 = arith.constant dense<0xFF800000> : tensor<128xf32>
  40. %c0_i32 = arith.constant 0 : i32
  41. %cst_1 = arith.constant dense<0> : tensor<128x64xi32>
  42. %cst_2 = arith.constant dense<0.000000e+00> : tensor<128x128xf16>
  43. %c1_i32 = arith.constant 1 : i32
  44. %cst_3 = arith.constant dense<0.000000e+00> : tensor<128x128xf32>
  45. %c63_i32 = arith.constant 63 : i32
  46. %c64_i32 = arith.constant 64 : i32
  47. %c127_i32 = arith.constant 127 : i32
  48. %c128_i32 = arith.constant 128 : i32
  49. %c24_i64 = arith.constant 24 : i64
  50. %0 = tt.get_program_id x : i32
  51. %1 = tt.get_program_id z : i32
  52. %2 = arith.extsi %1 : i32 to i64
  53. %3 = tt.get_program_id y : i32
  54. %4 = arith.extsi %3 : i32 to i64
  55. %5 = arith.muli %2, %c24_i64 : i64
  56. %6 = arith.addi %5, %4 : i64
  57. %7 = arith.addi %arg18, %c127_i32 : i32
  58. %8 = arith.divsi %7, %c128_i32 : i32
  59. %9 = arith.extsi %8 : i32 to i64
  60. %10 = arith.muli %6, %9 : i64
  61. %11 = arith.addi %arg19, %c63_i32 : i32
  62. %12 = arith.divsi %11, %c64_i32 : i32
  63. %13 = arith.extsi %12 : i32 to i64
  64. %14 = arith.muli %6, %13 : i64
  65. %15 = arith.muli %0, %c128_i32 : i32
  66. %16 = tt.make_range {end = 128 : i32, start = 0 : i32} : tensor<128xi32>
  67. %17 = tt.splat %15 : i32 -> tensor<128xi32>
  68. %18 = arith.addi %17, %16 : tensor<128xi32>
  69. %19 = tt.make_range {end = 64 : i32, start = 0 : i32} : tensor<64xi32>
  70. %20 = arith.extsi %arg6 : i32 to i64
  71. %21 = arith.muli %2, %20 : i64
  72. %22 = arith.extsi %arg7 : i32 to i64
  73. %23 = arith.muli %4, %22 : i64
  74. %24 = arith.addi %21, %23 : i64
  75. %25 = tt.addptr %arg0, %24 : !tt.ptr<i8>, i64
  76. %26 = tt.expand_dims %18 {axis = 1 : i32} : tensor<128xi32> -> tensor<128x1xi32>
  77. %27 = tt.splat %arg8 : i32 -> tensor<128x1xi32>
  78. %28 = arith.muli %26, %27 : tensor<128x1xi32>
  79. %29 = tt.splat %25 : !tt.ptr<i8> -> tensor<128x1x!tt.ptr<i8>>
  80. %30 = tt.addptr %29, %28 : tensor<128x1x!tt.ptr<i8>>, tensor<128x1xi32>
  81. %31 = tt.expand_dims %16 {axis = 0 : i32} : tensor<128xi32> -> tensor<1x128xi32>
  82. %32 = tt.broadcast %30 : tensor<128x1x!tt.ptr<i8>> -> tensor<128x128x!tt.ptr<i8>>
  83. %33 = tt.broadcast %31 : tensor<1x128xi32> -> tensor<128x128xi32>
  84. %34 = tt.addptr %32, %33 : tensor<128x128x!tt.ptr<i8>>, tensor<128x128xi32>
  85. %35 = tt.addptr %arg3, %10 : !tt.ptr<f32>, i64
  86. %36 = tt.addptr %35, %0 : !tt.ptr<f32>, i32
  87. %37 = arith.extsi %arg9 : i32 to i64
  88. %38 = arith.muli %2, %37 : i64
  89. %39 = arith.extsi %arg10 : i32 to i64
  90. %40 = arith.muli %4, %39 : i64
  91. %41 = arith.addi %38, %40 : i64
  92. %42 = tt.addptr %arg1, %41 : !tt.ptr<i8>, i64
  93. %43 = tt.expand_dims %19 {axis = 0 : i32} : tensor<64xi32> -> tensor<1x64xi32>
  94. %44 = tt.splat %arg11 : i32 -> tensor<1x64xi32>
  95. %45 = arith.muli %43, %44 : tensor<1x64xi32>
  96. %46 = tt.splat %42 : !tt.ptr<i8> -> tensor<1x64x!tt.ptr<i8>>
  97. %47 = tt.addptr %46, %45 : tensor<1x64x!tt.ptr<i8>>, tensor<1x64xi32>
  98. %48 = tt.expand_dims %16 {axis = 1 : i32} : tensor<128xi32> -> tensor<128x1xi32>
  99. %49 = tt.broadcast %47 : tensor<1x64x!tt.ptr<i8>> -> tensor<128x64x!tt.ptr<i8>>
  100. %50 = tt.broadcast %48 : tensor<128x1xi32> -> tensor<128x64xi32>
  101. %51 = tt.addptr %49, %50 : tensor<128x64x!tt.ptr<i8>>, tensor<128x64xi32>
  102. %52 = tt.addptr %arg4, %14 : !tt.ptr<f32>, i64
  103. %53 = arith.extsi %arg12 : i32 to i64
  104. %54 = arith.muli %2, %53 : i64
  105. %55 = arith.extsi %arg13 : i32 to i64
  106. %56 = arith.muli %4, %55 : i64
  107. %57 = arith.addi %54, %56 : i64
  108. %58 = tt.addptr %arg2, %57 : !tt.ptr<f16>, i64
  109. %59 = tt.expand_dims %19 {axis = 1 : i32} : tensor<64xi32> -> tensor<64x1xi32>
  110. %60 = tt.splat %arg14 : i32 -> tensor<64x1xi32>
  111. %61 = arith.muli %59, %60 : tensor<64x1xi32>
  112. %62 = tt.splat %58 : !tt.ptr<f16> -> tensor<64x1x!tt.ptr<f16>>
  113. %63 = tt.addptr %62, %61 : tensor<64x1x!tt.ptr<f16>>, tensor<64x1xi32>
  114. %64 = tt.broadcast %63 : tensor<64x1x!tt.ptr<f16>> -> tensor<64x128x!tt.ptr<f16>>
  115. %65 = tt.broadcast %31 : tensor<1x128xi32> -> tensor<64x128xi32>
  116. %66 = tt.addptr %64, %65 : tensor<64x128x!tt.ptr<f16>>, tensor<64x128xi32>
  117. %67 = arith.extsi %arg15 : i32 to i64
  118. %68 = arith.muli %2, %67 : i64
  119. %69 = arith.extsi %arg16 : i32 to i64
  120. %70 = arith.muli %4, %69 : i64
  121. %71 = arith.addi %68, %70 : i64
  122. %72 = tt.addptr %arg5, %71 : !tt.ptr<bf16>, i64
  123. %73 = tt.splat %arg17 : i32 -> tensor<128x1xi32>
  124. %74 = arith.muli %26, %73 : tensor<128x1xi32>
  125. %75 = tt.splat %72 : !tt.ptr<bf16> -> tensor<128x1x!tt.ptr<bf16>>
  126. %76 = tt.addptr %75, %74 : tensor<128x1x!tt.ptr<bf16>>, tensor<128x1xi32>
  127. %77 = tt.broadcast %76 : tensor<128x1x!tt.ptr<bf16>> -> tensor<128x128x!tt.ptr<bf16>>
  128. %78 = tt.addptr %77, %33 : tensor<128x128x!tt.ptr<bf16>>, tensor<128x128xi32>
  129. %79 = tt.splat %arg18 : i32 -> tensor<128x1xi32>
  130. %80 = arith.cmpi slt, %26, %79 : tensor<128x1xi32>
  131. %81 = tt.broadcast %80 : tensor<128x1xi1> -> tensor<128x128xi1>
  132. %82 = tt.load %34, %81 : tensor<128x128x!tt.ptr<i8>>
  133. %83 = tt.load %36 : !tt.ptr<f32>
  134. %84:6 = scf.for %arg20 = %c0_i32 to %arg19 step %c64_i32 iter_args(%arg21 = %cst, %arg22 = %cst_3, %arg23 = %cst_0, %arg24 = %51, %arg25 = %52, %arg26 = %66) -> (tensor<128xf32>, tensor<128x128xf32>, tensor<128xf32>, tensor<128x64x!tt.ptr<i8>>, !tt.ptr<f32>, tensor<64x128x!tt.ptr<f16>>) : i32 {
  135. %89 = arith.subi %arg19, %arg20 : i32
  136. %90 = tt.splat %89 : i32 -> tensor<1x64xi32>
  137. %91 = arith.cmpi slt, %43, %90 : tensor<1x64xi32>
  138. %92 = tt.broadcast %91 : tensor<1x64xi1> -> tensor<128x64xi1>
  139. %93 = tt.load %arg24, %92 : tensor<128x64x!tt.ptr<i8>>
  140. %94 = tt.load %arg25 : !tt.ptr<f32>
  141. %95 = tt.dot %82, %93, %cst_1, inputPrecision = tf32 : tensor<128x128xi8> * tensor<128x64xi8> -> tensor<128x64xi32>
  142. %96 = arith.sitofp %95 : tensor<128x64xi32> to tensor<128x64xf32>
  143. %97 = tt.splat %83 : f32 -> tensor<128x64xf32>
  144. %98 = arith.mulf %96, %97 : tensor<128x64xf32>
  145. %99 = tt.splat %94 : f32 -> tensor<128x64xf32>
  146. %100 = arith.mulf %98, %99 : tensor<128x64xf32>
  147. %101 = "tt.reduce"(%100) <{axis = 1 : i32}> ({
  148. ^bb0(%arg27: f32, %arg28: f32):
  149. %130 = arith.maxnumf %arg27, %arg28 : f32
  150. tt.reduce.return %130 : f32
  151. }) : (tensor<128x64xf32>) -> tensor<128xf32>
  152. %102 = arith.maxnumf %arg23, %101 : tensor<128xf32>
  153. %103 = tt.expand_dims %102 {axis = 1 : i32} : tensor<128xf32> -> tensor<128x1xf32>
  154. %104 = tt.broadcast %103 : tensor<128x1xf32> -> tensor<128x64xf32>
  155. %105 = arith.subf %100, %104 : tensor<128x64xf32>
  156. %106 = math.exp2 %105 : tensor<128x64xf32>
  157. %107 = "tt.reduce"(%106) <{axis = 1 : i32}> ({
  158. ^bb0(%arg27: f32, %arg28: f32):
  159. %130 = arith.addf %arg27, %arg28 : f32
  160. tt.reduce.return %130 : f32
  161. }) : (tensor<128x64xf32>) -> tensor<128xf32>
  162. %108 = arith.subf %arg23, %102 : tensor<128xf32>
  163. %109 = math.exp2 %108 : tensor<128xf32>
  164. %110 = arith.mulf %arg21, %109 : tensor<128xf32>
  165. %111 = arith.addf %110, %107 : tensor<128xf32>
  166. %112 = tt.expand_dims %109 {axis = 1 : i32} : tensor<128xf32> -> tensor<128x1xf32>
  167. %113 = tt.broadcast %112 : tensor<128x1xf32> -> tensor<128x128xf32>
  168. %114 = arith.mulf %arg22, %113 : tensor<128x128xf32>
  169. %115 = tt.splat %89 : i32 -> tensor<64x1xi32>
  170. %116 = arith.cmpi slt, %59, %115 : tensor<64x1xi32>
  171. %117 = tt.broadcast %116 : tensor<64x1xi1> -> tensor<64x128xi1>
  172. %118 = tt.load %arg26, %117 : tensor<64x128x!tt.ptr<f16>>
  173. %119 = arith.truncf %106 : tensor<128x64xf32> to tensor<128x64xf16>
  174. %120 = tt.dot %119, %118, %cst_2, inputPrecision = tf32 : tensor<128x64xf16> * tensor<64x128xf16> -> tensor<128x128xf16>
  175. %121 = arith.extf %120 : tensor<128x128xf16> to tensor<128x128xf32>
  176. %122 = arith.addf %114, %121 : tensor<128x128xf32>
  177. %123 = arith.muli %arg11, %c64_i32 : i32
  178. %124 = tt.splat %123 : i32 -> tensor<128x64xi32>
  179. %125 = tt.addptr %arg24, %124 : tensor<128x64x!tt.ptr<i8>>, tensor<128x64xi32>
  180. %126 = tt.addptr %arg25, %c1_i32 : !tt.ptr<f32>, i32
  181. %127 = arith.muli %arg14, %c64_i32 : i32
  182. %128 = tt.splat %127 : i32 -> tensor<64x128xi32>
  183. %129 = tt.addptr %arg26, %128 : tensor<64x128x!tt.ptr<f16>>, tensor<64x128xi32>
  184. scf.yield %111, %122, %102, %125, %126, %129 : tensor<128xf32>, tensor<128x128xf32>, tensor<128xf32>, tensor<128x64x!tt.ptr<i8>>, !tt.ptr<f32>, tensor<64x128x!tt.ptr<f16>>
  185. } {tt.divisibility_arg1 = dense<64> : tensor<1xi32>}
  186. %85 = tt.expand_dims %84#0 {axis = 1 : i32} : tensor<128xf32> -> tensor<128x1xf32>
  187. %86 = tt.broadcast %85 : tensor<128x1xf32> -> tensor<128x128xf32>
  188. %87 = arith.divf %84#1, %86 : tensor<128x128xf32>
  189. %88 = arith.truncf %87 : tensor<128x128xf32> to tensor<128x128xbf16>
  190. tt.store %78, %88, %81 : tensor<128x128x!tt.ptr<bf16>>
  191. tt.return
  192. }
  193. }
  194.  
  195. {-#
  196. external_resources: {
  197. mlir_reproducer: {
  198. pipeline: "builtin.module(convert-triton-to-tritongpu{num-ctas=1 num-warps=8 target=cuda:75 threads-per-warp=32}, tritongpu-coalesce, triton-nvidia-gpu-plan-cta, tritongpu-remove-layout-conversions, tritongpu-optimize-thread-locality, tritongpu-accelerate-matmul, tritongpu-remove-layout-conversions, tritongpu-optimize-dot-operands{hoist-layout-conversion=false}, cse, loop-invariant-code-motion, tritongpu-prefetch, tritongpu-optimize-dot-operands{hoist-layout-conversion=false}, tritongpu-coalesce-async-copy, tritongpu-remove-layout-conversions, tritongpu-reduce-data-duplication, tritongpu-reorder-instructions, cse, symbol-dce, canonicalize{ max-iterations=10 max-num-rewrites=-1 region-simplify=normal test-convergence=false top-down=true})",
  199. disable_threading: false,
  200. verify_each: true
  201. }
  202. }
  203. #-}
  204. /usr/local/lib/python3.10/dist-packages/sageattention/attn_qk_int8_per_block.py:40:0: error: Failures have been detected while processing an MLIR pass pipeline
  205. /usr/local/lib/python3.10/dist-packages/sageattention/attn_qk_int8_per_block.py:40:0: note: Pipeline failed while executing [`TritonGPUAccelerateMatmul` on 'builtin.module' operation]: reproducer generated at `std::errs, please share the reproducer above with Triton project.`
  206. 0% 0/25 [00:06<?, ?it/s]
  207. Traceback (most recent call last):
  208. File "/content/FramePack/demo_gradio.py", line 241, in worker
  209. generated_latents = sample_hunyuan(
  210. File "/usr/local/lib/python3.10/dist-packages/torch/utils/_contextlib.py", line 116, in decorate_context
  211. return func(*args, **kwargs)
  212. File "/content/FramePack/diffusers_helper/pipelines/k_diffusion_hunyuan.py", line 116, in sample_hunyuan
  213. results = sample_unipc(k_model, latents, sigmas, extra_args=sampler_kwargs, disable=False, callback=callback)
  214. File "/content/FramePack/diffusers_helper/k_diffusion/uni_pc_fm.py", line 141, in sample_unipc
  215. return FlowMatchUniPC(model, extra_args=extra_args, variant=variant).sample(noise, sigmas=sigmas, callback=callback, disable_pbar=disable)
  216. File "/content/FramePack/diffusers_helper/k_diffusion/uni_pc_fm.py", line 118, in sample
  217. model_prev_list = [self.model_fn(x, vec_t)]
  218. File "/content/FramePack/diffusers_helper/k_diffusion/uni_pc_fm.py", line 23, in model_fn
  219. return self.model(x, t, **self.extra_args)
  220. File "/content/FramePack/diffusers_helper/k_diffusion/wrapper.py", line 37, in k_model
  221. pred_positive = transformer(hidden_states=hidden_states, timestep=timestep, return_dict=False, **extra_args['positive'])[0].float()
  222. File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl
  223. return self._call_impl(*args, **kwargs)
  224. File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1762, in _call_impl
  225. return forward_call(*args, **kwargs)
  226. File "/content/FramePack/diffusers_helper/models/hunyuan_video_packed.py", line 973, in forward
  227. hidden_states, encoder_hidden_states = self.gradient_checkpointing_method(
  228. File "/content/FramePack/diffusers_helper/models/hunyuan_video_packed.py", line 832, in gradient_checkpointing_method
  229. result = block(*args)
  230. File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl
  231. return self._call_impl(*args, **kwargs)
  232. File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1762, in _call_impl
  233. return forward_call(*args, **kwargs)
  234. File "/content/FramePack/diffusers_helper/models/hunyuan_video_packed.py", line 652, in forward
  235. attn_output, context_attn_output = self.attn(
  236. File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl
  237. return self._call_impl(*args, **kwargs)
  238. File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1762, in _call_impl
  239. return forward_call(*args, **kwargs)
  240. File "/usr/local/lib/python3.10/dist-packages/diffusers/models/attention_processor.py", line 605, in forward
  241. return self.processor(
  242. File "/content/FramePack/diffusers_helper/models/hunyuan_video_packed.py", line 172, in __call__
  243. hidden_states = attn_varlen_func(query, key, value, cu_seqlens_q, cu_seqlens_kv, max_seqlen_q, max_seqlen_kv)
  244. File "/content/FramePack/diffusers_helper/models/hunyuan_video_packed.py", line 111, in attn_varlen_func
  245. x = sageattn(q, k, v, tensor_layout='NHD')
  246. File "/usr/local/lib/python3.10/dist-packages/sageattention/core.py", line 110, in sageattn
  247. o = attn_false(q_int8, k_int8, v, q_scale, k_scale, tensor_layout=tensor_layout, output_dtype=dtype)
  248. File "/usr/local/lib/python3.10/dist-packages/sageattention/attn_qk_int8_per_block.py", line 113, in forward
  249. _attn_fwd[grid](
  250. File "/usr/local/lib/python3.10/dist-packages/triton/runtime/jit.py", line 347, in <lambda>
  251. return lambda *args, **kwargs: self.run(grid=grid, warmup=False, *args, **kwargs)
  252. File "/usr/local/lib/python3.10/dist-packages/triton/runtime/jit.py", line 569, in run
  253. kernel = self.compile(src, target=target, options=options.__dict__)
  254. File "/usr/local/lib/python3.10/dist-packages/triton/compiler/compiler.py", line 284, in compile
  255. next_module = compile_ir(module, metadata)
  256. File "/usr/local/lib/python3.10/dist-packages/triton/backends/nvidia/compiler.py", line 449, in <lambda>
  257. stages["ttgir"] = lambda src, metadata: self.make_ttgir(src, metadata, options, capability)
  258. File "/usr/local/lib/python3.10/dist-packages/triton/backends/nvidia/compiler.py", line 312, in make_ttgir
  259. pm.run(mod)
  260. RuntimeError: PassManager::run failed
  261. Unloaded DynamicSwap_LlamaModel as complete.
  262. Unloaded CLIPTextModel as complete.
  263. Unloaded SiglipVisionModel as complete.
  264. Unloaded AutoencoderKLHunyuanVideo as complete.
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement
Public Pastes
Advertisement
We use cookies for various purposes including analytics. By continuing to use Pastebin, you agree to our use of cookies as described in the Cookies Policy.  OK, I Understand
Not a member of Pastebin yet?
Sign Up, it unlocks many cool features!