collabora
Installation issue: tokenizers version
My system: Ubuntu 22.04 4 x 4090 +---------------------------------------------------------------------------------------+ | NVIDIA-SMI 535.183.01 Driver Version: 535.183.01 CUDA Version: 12.2 | |-----------------------------------------+----------------------+----------------------+ | GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. | | | | MIG M. | |=========================================+======================+======================| | 0 NVIDIA GeForce RTX 4090 Off | 00000000:18:00.0 Off | Off | | 0% 45C P8 31W / 450W | 11MiB / 24564MiB | 0% Default | | | | N/A | +-----------------------------------------+----------------------+----------------------+ | 1 NVIDIA GeForce RTX 4090 Off | 00000000:51:00.0 On | Off | | 0% 44C P8 32W / 450W | 52MiB / 24564MiB | 5% Default | | | | N/A | +-----------------------------------------+----------------------+----------------------+ | 2 NVIDIA GeForce RTX 4090 Off | 00000000:8A:00.0 Off | Off | | 0% 40C P8 26W / 450W | 628MiB / 24564MiB | 8% Default | | | | N/A | +-----------------------------------------+----------------------+----------------------+ | 3 NVIDIA GeForce RTX 4090 Off | 00000000:C3:00.0 Off | Off | | 0% 43C P8 35W / 450W | 11MiB / 24564MiB | 0% Default | | | | N/A | +-----------------------------------------+----------------------+----------------------+
Following the build and run with docker compose, I get this module versioning error during docker compose build:
280.3 Traceback (most recent call last):
280.3 File "<string>", line 1, in <module>
280.3 File "/usr/local/lib/python3.10/dist-packages/transformers/__init__.py", line 26, in <module>
280.3 from . import dependency_versions_check
280.3 File "/usr/local/lib/python3.10/dist-packages/transformers/dependency_versions_check.py", line 57, in <module>
280.3 require_version_core(deps[pkg])
280.3 File "/usr/local/lib/python3.10/dist-packages/transformers/utils/versions.py", line 117, in require_version_core
280.3 return require_version(requirement, hint)
280.3 File "/usr/local/lib/python3.10/dist-packages/transformers/utils/versions.py", line 111, in require_version
280.3 _compare_versions(op, got_ver, want_ver, requirement, pkg, hint)
280.3 File "/usr/local/lib/python3.10/dist-packages/transformers/utils/versions.py", line 44, in _compare_versions
280.3 raise ImportError(
280.3 ImportError: tokenizers>=0.19,<0.20 is required for a normal functioning of this module, but found tokenizers==0.20.0.
280.3 Try: `pip install transformers -U` or `pip install -e '.[dev]'` if you're working with git main
------
failed to solve: process "/bin/sh -c ./setup-whisperfusion.sh" did not complete successfully: exit code: 1
So I edited setup-whisperfusion.sh from:
pip install -U huggingface_hub tokenizers
to:
pip install -U huggingface_hub tokenizers==0.19.0
Now getting the following error after running docker compose up:
dssadmin@gpu022:~/WhisperFusion$ docker compose up
[+] Running 8/8
✔ nginx Pulled 6.6s
✔ e4fff0779e6d Pull complete 4.5s
✔ 2a0cb278fd9f Pull complete 5.7s
✔ 7045d6c32ae2 Pull complete 5.8s
✔ 03de31afb035 Pull complete 5.8s
✔ 0f17be8dcff2 Pull complete 5.8s
✔ 14b7e5e8f394 Pull complete 5.8s
✔ 23fa5a7b99a6 Pull complete 5.8s
[+] Running 3/2
✔ Network whisperfusion_default Created 0.1s
✔ Container whisperfusion-whisperfusion-1 Created 0.4s
✔ Container whisperfusion-nginx-1 Created 0.1s
Attaching to nginx-1, whisperfusion-1
whisperfusion-1 | Running build-models.sh...
whisperfusion-1 | whisper_small_en directory does not exist or is empty. Running build-whisper.sh...
whisperfusion-1 | Downloading PyTorch weights for small.en model
nginx-1 | Starting web service with Nginx
whisperfusion-1 | Building Whisper TensorRT Engine...
whisperfusion-1 | [TensorRT-LLM] TensorRT-LLM version: 0.10.0
whisperfusion-1 | [09/03/2024-16:41:34] [TRT-LLM] [I] plugin_arg is None, setting it as float16 automatically.
whisperfusion-1 | [09/03/2024-16:41:34] [TRT-LLM] [I] plugin_arg is None, setting it as float16 automatically.
whisperfusion-1 | [09/03/2024-16:41:34] [TRT-LLM] [I] plugin_arg is None, setting it as float16 automatically.
whisperfusion-1 | [09/03/2024-16:41:34] [TRT] [W] Unable to determine GPU memory usage: forward compatibility was attempted on non supported HW
whisperfusion-1 | [09/03/2024-16:41:34] [TRT] [W] Unable to determine GPU memory usage: forward compatibility was attempted on non supported HW
whisperfusion-1 | [09/03/2024-16:41:34] [TRT] [I] [MemUsageChange] Init CUDA: CPU +0, GPU +0, now: CPU 576, GPU 0 (MiB)
whisperfusion-1 | terminate called without an active exception
whisperfusion-1 | [e4b7b7abb3d0:00018] *** Process received signal ***
whisperfusion-1 | [e4b7b7abb3d0:00018] Signal: Aborted (6)
whisperfusion-1 | [e4b7b7abb3d0:00018] Signal code: (-6)
whisperfusion-1 | [e4b7b7abb3d0:00018] [ 0] /usr/lib/x86_64-linux-gnu/libc.so.6(+0x42520)[0x7f0f29b0f520]
whisperfusion-1 | [e4b7b7abb3d0:00018] [ 1] /usr/lib/x86_64-linux-gnu/libc.so.6(pthread_kill+0x12c)[0x7f0f29b639fc]
whisperfusion-1 | [e4b7b7abb3d0:00018] [ 2] /usr/lib/x86_64-linux-gnu/libc.so.6(raise+0x16)[0x7f0f29b0f476]
whisperfusion-1 | [e4b7b7abb3d0:00018] [ 3] /usr/lib/x86_64-linux-gnu/libc.so.6(abort+0xd3)[0x7f0f29af57f3]
whisperfusion-1 | [e4b7b7abb3d0:00018] [ 4] /usr/lib/x86_64-linux-gnu/libstdc++.so.6(+0xa2b9e)[0x7f0f28676b9e]
whisperfusion-1 | [e4b7b7abb3d0:00018] [ 5] /usr/lib/x86_64-linux-gnu/libstdc++.so.6(+0xae20c)[0x7f0f2868220c]
whisperfusion-1 | [e4b7b7abb3d0:00018] [ 6] /usr/lib/x86_64-linux-gnu/libstdc++.so.6(+0xae277)[0x7f0f28682277]
whisperfusion-1 | [e4b7b7abb3d0:00018] [ 7] /usr/local/lib/python3.10/dist-packages/tensorrt_libs/libnvinfer.so.10(+0x4d1d47)[0x7f0e2d21cd47]
whisperfusion-1 | [e4b7b7abb3d0:00018] [ 8] /usr/local/lib/python3.10/dist-packages/tensorrt_bindings/tensorrt.so(+0x87b58)[0x7f0d83c87b58]
whisperfusion-1 | [e4b7b7abb3d0:00018] [ 9] /usr/local/lib/python3.10/dist-packages/tensorrt_bindings/tensorrt.so(+0x46123)[0x7f0d83c46123]
whisperfusion-1 | [e4b7b7abb3d0:00018] [10] python3(+0x15adae)[0x563154f02dae]
whisperfusion-1 | [e4b7b7abb3d0:00018] [11] python3(_PyObject_MakeTpCall+0x25b)[0x563154ef952b]
whisperfusion-1 | [e4b7b7abb3d0:00018] [12] python3(+0x169680)[0x563154f11680]
whisperfusion-1 | [e4b7b7abb3d0:00018] [13] python3(+0x165dc7)[0x563154f0ddc7]
whisperfusion-1 | [e4b7b7abb3d0:00018] [14] python3(+0x1518db)[0x563154ef98db]
whisperfusion-1 | [e4b7b7abb3d0:00018] [15] /usr/local/lib/python3.10/dist-packages/tensorrt_bindings/tensorrt.so(+0x41b73)[0x7f0d83c41b73]
whisperfusion-1 | [e4b7b7abb3d0:00018] [16] python3(_PyObject_MakeTpCall+0x25b)[0x563154ef952b]
whisperfusion-1 | [e4b7b7abb3d0:00018] [17] python3(_PyEval_EvalFrameDefault+0x6f0b)[0x563154ef216b]
whisperfusion-1 | [e4b7b7abb3d0:00018] [18] python3(_PyObject_FastCallDictTstate+0xc4)[0x563154ef86c4]
whisperfusion-1 | [e4b7b7abb3d0:00018] [19] python3(+0x1657a4)[0x563154f0d7a4]
whisperfusion-1 | [e4b7b7abb3d0:00018] [20] python3(_PyObject_MakeTpCall+0x1fc)[0x563154ef94cc]
whisperfusion-1 | [e4b7b7abb3d0:00018] [21] python3(_PyEval_EvalFrameDefault+0x68bc)[0x563154ef1b1c]
whisperfusion-1 | [e4b7b7abb3d0:00018] [22] python3(_PyFunction_Vectorcall+0x7c)[0x563154f036ac]
whisperfusion-1 | [e4b7b7abb3d0:00018] [23] python3(_PyEval_EvalFrameDefault+0x6d5)[0x563154eeb935]
whisperfusion-1 | [e4b7b7abb3d0:00018] [24] python3(_PyFunction_Vectorcall+0x7c)[0x563154f036ac]
whisperfusion-1 | [e4b7b7abb3d0:00018] [25] python3(_PyEval_EvalFrameDefault+0x6d5)[0x563154eeb935]
whisperfusion-1 | [e4b7b7abb3d0:00018] [26] python3(+0x140096)[0x563154ee8096]
whisperfusion-1 | [e4b7b7abb3d0:00018] [27] python3(PyEval_EvalCode+0x86)[0x563154fddf66]
whisperfusion-1 | [e4b7b7abb3d0:00018] [28] python3(+0x260e98)[0x563155008e98]
whisperfusion-1 | [e4b7b7abb3d0:00018] [29] python3(+0x25a79b)[0x56315500279b]
whisperfusion-1 | [e4b7b7abb3d0:00018] *** End of error message ***
Not sure what to do from here.
Ok.. on the host machine I have Cuda 12.2, not 12.4. So I changed it in the Dockerfile to: 'RG BASE_TAG=12.2.0-runtime-ubuntu22.04'.