search

siliconflow / onediff

OneDiff: An out-of-the-box acceleration library for diffusion models.
https://github.com/siliconflow/onediff/wiki
Apache License 2.0
1.71k stars 105 forks source link

[Bug] Support to compile CogVideoXPipeline #1099

Open loretoparisi opened 2 months ago

loretoparisi commented 2 months ago

Your current environment information

Collecting environment information...
PyTorch version: 2.4.0+cu121
Is debug build: False
CUDA used to build PyTorch: 12.1
ROCM used to build PyTorch: N/A

OneFlow version: none
Nexfort version: none
OneDiff version: none
OneDiffX version: none

OS: Ubuntu 20.04.6 LTS (x86_64)
GCC version: (Ubuntu 9.4.0-1ubuntu1~20.04.2) 9.4.0
Clang version: Could not collect
CMake version: Could not collect
Libc version: glibc-2.31

Python version: 3.8.10 (default, Nov 22 2023, 10:22:35)  [GCC 9.4.0] (64-bit runtime)
Python platform: Linux-5.10.219-208.866.amzn2.x86_64-x86_64-with-glibc2.29
Is CUDA available: True
CUDA runtime version: Could not collect
CUDA_MODULE_LOADING set to: LAZY
GPU models and configuration: 
GPU 0: NVIDIA A10G
GPU 1: NVIDIA A10G
GPU 2: NVIDIA A10G
GPU 3: NVIDIA A10G

Nvidia driver version: 535.183.01
cuDNN version: Probably one of the following:
/usr/lib/x86_64-linux-gnu/libcudnn.so.8.2.4
/usr/lib/x86_64-linux-gnu/libcudnn_adv_infer.so.8.2.4
/usr/lib/x86_64-linux-gnu/libcudnn_adv_train.so.8.2.4
/usr/lib/x86_64-linux-gnu/libcudnn_cnn_infer.so.8.2.4
/usr/lib/x86_64-linux-gnu/libcudnn_cnn_train.so.8.2.4
/usr/lib/x86_64-linux-gnu/libcudnn_ops_infer.so.8.2.4
/usr/lib/x86_64-linux-gnu/libcudnn_ops_train.so.8.2.4
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: True

CPU:
Architecture:                         x86_64
CPU op-mode(s):                       32-bit, 64-bit
Byte Order:                           Little Endian
Address sizes:                        48 bits physical, 48 bits virtual
CPU(s):                               48
On-line CPU(s) list:                  0-47
Thread(s) per core:                   2
Core(s) per socket:                   24
Socket(s):                            1
NUMA node(s):                         1
Vendor ID:                            AuthenticAMD
CPU family:                           23
Model:                                49
Model name:                           AMD EPYC 7R32
Stepping:                             0
CPU MHz:                              3096.929
BogoMIPS:                             5599.34
Hypervisor vendor:                    KVM
Virtualization type:                  full
L1d cache:                            768 KiB
L1i cache:                            768 KiB
L2 cache:                             12 MiB
L3 cache:                             96 MiB
NUMA node0 CPU(s):                    0-47
Vulnerability Gather data sampling:   Not affected
Vulnerability Itlb multihit:          Not affected
Vulnerability L1tf:                   Not affected
Vulnerability Mds:                    Not affected
Vulnerability Meltdown:               Not affected
Vulnerability Mmio stale data:        Not affected
Vulnerability Reg file data sampling: Not affected
Vulnerability Retbleed:               Mitigation; untrained return thunk; SMT enabled with STIBP protection
Vulnerability Spec rstack overflow:   Mitigation; safe RET
Vulnerability Spec store bypass:      Mitigation; Speculative Store Bypass disabled via prctl and seccomp
Vulnerability Spectre v1:             Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2:             Mitigation; Retpolines, IBPB conditional, STIBP always-on, RSB filling, PBRSB-eIBRS Not affected
Vulnerability Srbds:                  Not affected
Vulnerability Tsx async abort:        Not affected
Flags:                                fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc cpuid extd_apicid aperfmperf tsc_known_freq pni pclmulqdq ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm cmp_legacy cr8_legacy abm sse4a misalignsse 3dnowprefetch topoext ssbd ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 clzero xsaveerptr rdpru wbnoinvd arat npt nrip_save rdpid

Versions of relevant libraries:
[pip3] numpy==1.24.4
[pip3] onnx==1.14.1
[pip3] onnxruntime-gpu==1.14.1
[pip3] pytorch-lightning==1.9.5
[pip3] sentence-transformers==2.2.2
[pip3] torch==2.4.0
[pip3] torchao==0.3.1
[pip3] torchmetrics==1.4.1
[pip3] torchtune==0.2.1
[pip3] torchvision==0.19.0
[pip3] transformers==4.44.0
[pip3] triton==3.0.0
[conda] Could not collect

🐛 Describe the bug

When attempting to save a CogVideoXPipeline using the official demo script modified adding onediffx

installed as

pip install -r requirements.txt
# Added **experimental** support for onediff, this reduced sampling time by ~40% for me, reaching 4.23 s/it on 4090 with 49 frames. 
# This requires using Linux, torch 2.4.0, onediff and nexfort installation:
pip install --pre onediff onediffx
pip install nexfort

where the requiremets are

diffusers>=0.30.1 #git+https://github.com/huggingface/diffusers.git@main#egg=diffusers is suggested
transformers>=4.44.0  # The development team is working on version 0.44.2
accelerate>=0.33.0 #git+https://github.com/huggingface/accelerate.git@main#egg=accelerate is suggested
sentencepiece>=0.2.0 # T5 used
SwissArmyTransformer>=0.4.12
numpy
torch>=2.4.0 # Tested in 2.2 2.3 2.4 and 2.5, The development team is working on version 2.4.0.
torchvision>=0.19.0 # The development team is working on version 0.19.0.
gradio>=4.42.0 # For HF gradio demo
streamlit>=1.37.1 # For streamlit web demo
imageio==2.34.2 # For diffusers inference export video
imageio-ffmpeg==0.5.1 # For diffusers inference export video
openai>=1.42.0 # For prompt refiner
moviepy==1.0.3 # For export video
pillow==9.5.0
torchao==0.4.0

It will fail with an error:


# load pip
pipe = CogVideoXPipeline.from_pretrained(
            model_path,
            text_encoder=text_encoder,
            transformer=transformer,
            vae=vae,
            torch_dtype=dtype,
        ).to(device)
        pipe.scheduler = CogVideoXDPMScheduler.from_config(pipe.scheduler.config, timestep_spacing="trailing")

#  optmizations enabled
pipe.enable_model_cpu_offload(device=device)
pipe.enable_sequential_cpu_offload(device=device)
pipe.vae.enable_tiling()
pipe.vae.enable_slicing()

# generate
video = pipe(
        prompt=prompt,
        num_videos_per_prompt=num_videos_per_prompt,
        num_inference_steps=num_inference_steps,
        num_frames=num_frames,
        use_dynamic_cfg=True,  ## This id used for DPM Sechduler, for DDIM scheduler, it should be False
        guidance_scale=guidance_scale,
        generator=torch.Generator(device=device).manual_seed(42)
    ).frames[0]

 # save compiled pipeline if supported by compile_backend
if compile_backend == "onediff":
    save_pipe(pipe, dir="cached_pipe", overwrite=True)

The error was 'T5EncoderModel' object has no attribute '_deployable_module_dpl_graph', while the folder cached_pipe has been created but it is empty.

Additionally Torch Dynamo metrics here (at first run before the error)

I0830 11:29:18.335000 140500382687232 torch/_dynamo/utils.py:335] TorchDynamo compilation metrics:
I0830 11:29:18.335000 140500382687232 torch/_dynamo/utils.py:335] Function, Runtimes (s)
V0830 11:29:18.335000 140500382687232 torch/fx/experimental/symbolic_shapes.py:116] lru_cache_stats constrain_symbol_range: CacheInfo(hits=0, misses=0, maxsize=None, currsize=0)
V0830 11:29:18.335000 140500382687232 torch/fx/experimental/symbolic_shapes.py:116] lru_cache_stats evaluate_expr: CacheInfo(hits=0, misses=0, maxsize=256, currsize=0)
V0830 11:29:18.336000 140500382687232 torch/fx/experimental/symbolic_shapes.py:116] lru_cache_stats _simplify_floor_div: CacheInfo(hits=0, misses=0, maxsize=None, currsize=0)
V0830 11:29:18.336000 140500382687232 torch/fx/experimental/symbolic_shapes.py:116] lru_cache_stats _maybe_guard_rel: CacheInfo(hits=0, misses=0, maxsize=256, currsize=0)
V0830 11:29:18.336000 140500382687232 torch/fx/experimental/symbolic_shapes.py:116] lru_cache_stats _find: CacheInfo(hits=0, misses=0, maxsize=None, currsize=0)
V0830 11:29:18.336000 140500382687232 torch/fx/experimental/symbolic_shapes.py:116] lru_cache_stats has_hint: CacheInfo(hits=0, misses=0, maxsize=256, currsize=0)
V0830 11:29:18.336000 140500382687232 torch/fx/experimental/symbolic_shapes.py:116] lru_cache_stats size_hint: CacheInfo(hits=0, misses=0, maxsize=256, currsize=0)
V0830 11:29:18.336000 140500382687232 torch/fx/experimental/symbolic_shapes.py:116] lru_cache_stats simplify: CacheInfo(hits=0, misses=0, maxsize=None, currsize=0)
V0830 11:29:18.336000 140500382687232 torch/fx/experimental/symbolic_shapes.py:116] lru_cache_stats _update_divisible: CacheInfo(hits=0, misses=0, maxsize=None, currsize=0)
V0830 11:29:18.336000 140500382687232 torch/fx/experimental/symbolic_shapes.py:116] lru_cache_stats replace: CacheInfo(hits=0, misses=0, maxsize=None, currsize=0)
V0830 11:29:18.336000 140500382687232 torch/fx/experimental/symbolic_shapes.py:116] lru_cache_stats _maybe_evaluate_static: CacheInfo(hits=0, misses=0, maxsize=None, currsize=0)
V0830 11:29:18.336000 140500382687232 torch/fx/experimental/symbolic_shapes.py:116] lru_cache_stats get_implications: CacheInfo(hits=0, misses=0, maxsize=None, currsize=0)
V0830 11:29:18.336000 140500382687232 torch/fx/experimental/symbolic_shapes.py:116] lru_cache_stats get_axioms: CacheInfo(hits=0, misses=0, maxsize=None, currsize=0)
V0830 11:29:18.336000 140500382687232 torch/fx/experimental/symbolic_shapes.py:116] lru_cache_stats safe_expand: CacheInfo(hits=0, misses=0, maxsize=256, currsize=0)
V0830 11:29:18.337000 140500382687232 torch/fx/experimental/symbolic_shapes.py:116] lru_cache_stats uninteresting_files: CacheInfo(hits=0, misses=0, maxsize=None, currsize=0)
Shiroha-Key commented 2 months ago

I met completely same problem : 'T5EncoderModel' object has no attribute '_deployable_module_dpl_graph' same Additionally Torch Dynamo metrics

PyTorch version: 2.4.1+cu121
Is debug build: False
CUDA used to build PyTorch: 12.1
ROCM used to build PyTorch: N/A

OneFlow version: path: ['/root/miniforge3/envs/py310-CogVideo/lib/python3.10/site-packages/oneflow'], version: 0.9.1.dev20240923+cu121, git_commit: d23c061, cmake_build_type: Release, rdma: True, mlir: True, enterprise: False
Nexfort version: 0.1.dev271
OneDiff version: 1.2.0
OneDiffX version: 1.2.0

OS: Ubuntu 20.04.6 LTS (x86_64)
GCC version: (Ubuntu 9.4.0-1ubuntu1~20.04.2) 9.4.0
Clang version: Could not collect
CMake version: Could not collect
Libc version: glibc-2.31

Python version: 3.10.15 | packaged by conda-forge | (main, Sep 20 2024, 16:37:05) [GCC 13.3.0] (64-bit runtime)
Python platform: Linux-5.4.0-26-generic-x86_64-with-glibc2.31
Is CUDA available: True
CUDA runtime version: 12.1.66
CUDA_MODULE_LOADING set to: LAZY
GPU models and configuration: 
GPU 0: NVIDIA GeForce RTX 3090
GPU 1: NVIDIA GeForce RTX 3090
GPU 2: NVIDIA GeForce RTX 3090
GPU 3: NVIDIA GeForce RTX 3090
GPU 4: NVIDIA GeForce RTX 3090
GPU 5: NVIDIA GeForce RTX 3090
GPU 6: NVIDIA GeForce RTX 3090
GPU 7: NVIDIA GeForce RTX 3090

Nvidia driver version: 525.60.13
cuDNN version: Could not collect
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: True

CPU:
Architecture:                    x86_64
CPU op-mode(s):                  32-bit, 64-bit
Byte Order:                      Little Endian
Address sizes:                   46 bits physical, 57 bits virtual
CPU(s):                          96
On-line CPU(s) list:             0-95
Thread(s) per core:              2
Core(s) per socket:              24
Socket(s):                       2
NUMA node(s):                    2
Vendor ID:                       GenuineIntel
CPU family:                      6
Model:                           106
Model name:                      Intel(R) Xeon(R) Gold 6342 CPU @ 2.80GHz
Stepping:                        6
Frequency boost:                 enabled
CPU MHz:                         3373.456
CPU max MHz:                     3500.0000
CPU min MHz:                     800.0000
BogoMIPS:                        5600.00
Virtualization:                  VT-x
L1d cache:                       2.3 MiB
L1i cache:                       1.5 MiB
L2 cache:                        60 MiB
L3 cache:                        72 MiB
NUMA node0 CPU(s):               0-23,48-71
NUMA node1 CPU(s):               24-47,72-95
Vulnerability Itlb multihit:     Not affected
Vulnerability L1tf:              Not affected
Vulnerability Mds:               Not affected
Vulnerability Meltdown:          Not affected
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl and seccomp
Vulnerability Spectre v1:        Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2:        Mitigation; Enhanced IBRS, IBPB conditional, RSB filling
Vulnerability Tsx async abort:   Not affected
Flags:                           fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb cat_l3 invpcid_single ssbd mba ibrs ibpb stibp ibrs_enhanced tpr_shadow vnmi flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid cqm rdt_a avx512f avx512dq rdseed adx smap avx512ifma clflushopt clwb intel_pt avx512cd sha_ni avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local wbnoinvd dtherm ida arat pln pts avx512vbmi umip pku ospke avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg tme avx512_vpopcntdq rdpid md_clear pconfig flush_l1d arch_capabilities

Versions of relevant libraries:
[pip3] diffusers==0.30.3
[pip3] numpy==1.26.0
[pip3] torch==2.4.1
[pip3] torchvision==0.19.1
[pip3] transformers==4.44.2
[pip3] triton==3.0.0
[conda] numpy                     1.26.0                   pypi_0    pypi
[conda] torch                     2.4.1                    pypi_0    pypi
[conda] torchvision               0.19.1                   pypi_0    pypi
[conda] triton                    3.0.0                    pypi_0    pypi
Shiroha-Key commented 1 month ago

i noted there two line and add pipe.to("cuda")

# self.pipe.enable_model_cpu_offload()
# self.pipe.enable_sequential_cpu_offload()

the problem disappear @loretoparisi

loretoparisi commented 1 month ago

i noted there two line and add pipe.to("cuda")

# self.pipe.enable_model_cpu_offload()
# self.pipe.enable_sequential_cpu_offload()

the problem disappear @loretoparisi

Thanks so basically we have to disable cpu offloading before calling 

save_pipe(pipe, dir="cached_pipe", overwrite=True)

Are you still using VAE layer slicing and tiling?

pipe.vae.enable_tiling()
pipe.vae.enable_slicing()

and compiled successfully? Thanks!

Shiroha-Key commented 1 month ago

i tryed use your 'save_pipe(pipe, dir="cached_pipe", overwrite=True)' the problem has appeared again xd i suggest just dont use save_pipe and load_pipe for now.