xformers
pytorch-image-models
xformers | pytorch-image-models | |
---|---|---|
46 | 35 | |
7,631 | 29,828 | |
3.1% | 1.5% | |
9.3 | 9.4 | |
2 days ago | 5 days ago | |
Python | Python | |
GNU General Public License v3.0 or later | Apache License 2.0 |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
xformers
-
Animediff error
(venv) G:\A1111\Animediff\animatediff-cli-prompt-travel>animatediff generate -c config/prompts/01-ToonYou.json -W 256 -H 384 -L 128 -C 16 WARNING[XFORMERS]: xFormers can't load C++/CUDA extensions. xFormers was built for: PyTorch 2.1.0+cu121 with CUDA 1201 (you have 2.0.1+cpu) Python 3.10.11 (you have 3.10.6) Please reinstall xformers (see https://github.com/facebookresearch/xformers#installing-xformers) Memory-efficient attention, SwiGLU, sparse and more won't be available. Set XFORMERS_MORE_DETAILS=1 for more details G:\A1111\Animediff\animatediff-cli-prompt-travel\venv\lib\site-packages\torchvision\io\image.py:13: UserWarning: Failed to load image Python extension: 'Could not find module 'G:\A1111\Animediff\animatediff-cli-prompt-travel\venv\Lib\site-packages\torchvision\image.pyd' (or one of its dependencies). Try using the full path with constructor syntax.'If you don't plan on using image functionality from `torchvision.io`, you can ignore this warning. Otherwise, there might be something wrong with your environment. Did you have `libjpeg` or `libpng` installed before building `torchvision` from source? warn( 15:07:25 INFO Using generation config: config\prompts\01-ToonYou.json cli.py:291 ╭─────────────────────────────── Traceback (most recent call last) ────────────────────────────────╮ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\src\animatediff\cli.py:292 in generate │ │ │ │ 289 │ │ │ 290 │ config_path = config_path.absolute() │ │ 291 │ logger.info(f"Using generation config: {path_from_cwd(config_path)}") │ │ ❱ 292 │ model_config: ModelConfig = get_model_config(config_path) │ │ 293 │ is_v2 = is_v2_motion_module(data_dir.joinpath(model_config.motion_module)) │ │ 294 │ infer_config: InferenceConfig = get_infer_config(is_v2) │ │ 295 │ │ │ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\src\animatediff\settings.py:134 in │ │ get_model_config │ │ │ │ 131 │ │ 132 │ │ 133 def get_model_config(config_path: Path) -> ModelConfig: │ │ ❱ 134 │ settings = ModelConfig(json_config_path=config_path) │ │ 135 │ return settings │ │ 136 │ │ │ │ in pydantic.env_settings.BaseSettings.__init__:40 │ │ │ │ in pydantic.main.BaseModel.__init__:341 │ ╰──────────────────────────────────────────────────────────────────────────────────────────────────╯ ValidationError: 1 validation error for ModelConfig prompt extra fields not permitted (type=value_error.extra) (venv) G:\A1111\Animediff\animatediff-cli-prompt-travel>animatediff generate -c config/prompts/prompt1.json -W 256 -H 38 4 -L 128 -C 16 WARNING[XFORMERS]: xFormers can't load C++/CUDA extensions. xFormers was built for: PyTorch 2.1.0+cu121 with CUDA 1201 (you have 2.0.1+cpu) Python 3.10.11 (you have 3.10.6) Please reinstall xformers (see https://github.com/facebookresearch/xformers#installing-xformers) Memory-efficient attention, SwiGLU, sparse and more won't be available. Set XFORMERS_MORE_DETAILS=1 for more details G:\A1111\Animediff\animatediff-cli-prompt-travel\venv\lib\site-packages\torchvision\io\image.py:13: UserWarning: Failed to load image Python extension: 'Could not find module 'G:\A1111\Animediff\animatediff-cli-prompt-travel\venv\Lib\site-packages\torchvision\image.pyd' (or one of its dependencies). Try using the full path with constructor syntax.'If you don't plan on using image functionality from `torchvision.io`, you can ignore this warning. Otherwise, there might be something wrong with your environment. Did you have `libjpeg` or `libpng` installed before building `torchvision` from source? warn( 15:08:30 INFO Using generation config: config\prompts\prompt1.json cli.py:291 15:08:35 INFO is_v2=True util.py:361 INFO Using base model: runwayml\stable-diffusion-v1-5 cli.py:309 INFO Will save outputs to ./output\2023-10-30T15-08-35-epicrealism-epicrealism_naturalsinrc1vae cli.py:317 INFO Checking motion module... generate.py:331 INFO Loading tokenizer... generate.py:345 INFO Loading text encoder... generate.py:347 15:08:38 INFO Loading VAE... generate.py:349 INFO Loading UNet... generate.py:351 15:08:59 INFO Loaded 453.20928M-parameter motion module unet.py:578 15:09:00 INFO Using scheduler "euler_a" (EulerAncestralDiscreteScheduler) generate.py:363 INFO Loading weights from generate.py:368 G:\A1111\Animediff\animatediff-cli-prompt-travel\data\models\sd\epicrealism_naturalSin RC1VAE.safetensors ╭─────────────────────────────── Traceback (most recent call last) ────────────────────────────────╮ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\venv\lib\site-packages\urllib3\connection.py:17 │ │ 4 in _new_conn │ │ │ │ 171 │ │ │ extra_kw["socket_options"] = self.socket_options │ │ 172 │ │ │ │ 173 │ │ try: │ │ ❱ 174 │ │ │ conn = connection.create_connection( │ │ 175 │ │ │ │ (self._dns_host, self.port), self.timeout, **extra_kw │ │ 176 │ │ │ ) │ │ 177 │ │ │ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\venv\lib\site-packages\urllib3\util\connection. │ │ py:95 in create_connection │ │ │ │ 92 │ │ │ │ sock = None │ │ 93 │ │ │ 94 │ if err is not None: │ │ ❱ 95 │ │ raise err │ │ 96 │ │ │ 97 │ raise socket.error("getaddrinfo returns an empty list") │ │ 98 │ │ │ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\venv\lib\site-packages\urllib3\util\connection. │ │ py:85 in create_connection │ │ │ │ 82 │ │ │ │ sock.settimeout(timeout) │ │ 83 │ │ │ if source_address: │ │ 84 │ │ │ │ sock.bind(source_address) │ │ ❱ 85 │ │ │ sock.connect(sa) │ │ 86 │ │ │ return sock │ │ 87 │ │ │ │ 88 │ │ except socket.error as e: │ ╰──────────────────────────────────────────────────────────────────────────────────────────────────╯ TimeoutError: [WinError 10060] A connection attempt failed because the connected party did not properly respond after a period of time, or established connection failed because connected host has failed to respond During handling of the above exception, another exception occurred: ╭─────────────────────────────── Traceback (most recent call last) ────────────────────────────────╮ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\venv\lib\site-packages\urllib3\connectionpool.p │ │ y:703 in urlopen │ │ │ │ 700 │ │ │ │ self._prepare_proxy(conn) │ │ 701 │ │ │ │ │ 702 │ │ │ # Make the request on the httplib connection object. │ │ ❱ 703 │ │ │ httplib_response = self._make_request( │ │ 704 │ │ │ │ conn, │ │ 705 │ │ │ │ method, │ │ 706 │ │ │ │ url, │ │ │ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\venv\lib\site-packages\urllib3\connectionpool.p │ │ y:386 in _make_request │ │ │ │ 383 │ │ │ │ 384 │ │ # Trigger any extra validation we need to do. │ │ 385 │ │ try: │ │ ❱ 386 │ │ │ self._validate_conn(conn) │ │ 387 │ │ except (SocketTimeout, BaseSSLError) as e: │ │ 388 │ │ │ # Py2 raises this as a BaseSSLError, Py3 raises it as socket timeout. │ │ 389 │ │ │ self._raise_timeout(err=e, url=url, timeout_value=conn.timeout) │ │ │ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\venv\lib\site-packages\urllib3\connectionpool.p │ │ y:1042 in _validate_conn │ │ │ │ 1039 │ │ │ │ 1040 │ │ # Force connect early to allow us to validate the connection. │ │ 1041 │ │ if not getattr(conn, "sock", None): # AppEngine might not have `.sock` │ │ ❱ 1042 │ │ │ conn.connect() │ │ 1043 │ │ │ │ 1044 │ │ if not conn.is_verified: │ │ 1045 │ │ │ warnings.warn( │ │ │ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\venv\lib\site-packages\urllib3\connection.py:35 │ │ 8 in connect │ │ │ │ 355 │ │ │ 356 │ def connect(self): │ │ 357 │ │ # Add certificate verification │ │ ❱ 358 │ │ self.sock = conn = self._new_conn() │ │ 359 │ │ hostname = self.host │ │ 360 │ │ tls_in_tls = False │ │ 361 │ │ │ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\venv\lib\site-packages\urllib3\connection.py:17 │ │ 9 in _new_conn │ │ │ │ 176 │ │ │ ) │ │ 177 │ │ │ │ 178 │ │ except SocketTimeout: │ │ ❱ 179 │ │ │ raise ConnectTimeoutError( │ │ 180 │ │ │ │ self, │ │ 181 │ │ │ │ "Connection to %s timed out. (connect timeout=%s)" │ │ 182 │ │ │ │ % (self.host, self.timeout), │ ╰──────────────────────────────────────────────────────────────────────────────────────────────────╯ ConnectTimeoutError: (, 'Connection to raw.githubusercontent.com timed out. (connect timeout=None)') During handling of the above exception, another exception occurred: ╭─────────────────────────────── Traceback (most recent call last) ────────────────────────────────╮ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\venv\lib\site-packages\requests\adapters.py:489 │ │ in send │ │ │ │ 486 │ │ │ │ 487 │ │ try: │ │ 488 │ │ │ if not chunked: │ │ ❱ 489 │ │ │ │ resp = conn.urlopen( │ │ 490 │ │ │ │ │ method=request.method, │ │ 491 │ │ │ │ │ url=url, │ │ 492 │ │ │ │ │ body=request.body, │ │ │ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\venv\lib\site-packages\urllib3\connectionpool.p │ │ y:787 in urlopen │ │ │ │ 784 │ │ │ elif isinstance(e, (SocketError, HTTPException)): │ │ 785 │ │ │ │ e = ProtocolError("Connection aborted.", e) │ │ 786 │ │ │ │ │ ❱ 787 │ │ │ retries = retries.increment( │ │ 788 │ │ │ │ method, url, error=e, _pool=self, _stacktrace=sys.exc_info()[2] │ │ 789 │ │ │ ) │ │ 790 │ │ │ retries.sleep() │ │ │ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\venv\lib\site-packages\urllib3\util\retry.py:59 │ │ 2 in increment │ │ │ │ 589 │ │ ) │ │ 590 │ │ │ │ 591 │ │ if new_retry.is_exhausted(): │ │ ❱ 592 │ │ │ raise MaxRetryError(_pool, url, error or ResponseError(cause)) │ │ 593 │ │ │ │ 594 │ │ log.debug("Incremented Retry for (url='%s'): %r", url, new_retry) │ │ 595 │ ╰──────────────────────────────────────────────────────────────────────────────────────────────────╯ MaxRetryError: HTTPSConnectionPool(host='raw.githubusercontent.com', port=443): Max retries exceeded with url: /CompVis/stable-diffusion/main/configs/stable-diffusion/v1-inference.yaml (Caused by ConnectTimeoutError(, 'Connection to raw.githubusercontent.com timed out. (connect timeout=None)')) During handling of the above exception, another exception occurred: ╭─────────────────────────────── Traceback (most recent call last) ────────────────────────────────╮ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\src\animatediff\cli.py:326 in generate │ │ │ │ 323 │ global g_pipeline │ │ 324 │ global last_model_path │ │ 325 │ if g_pipeline is None or last_model_path != model_config.path.resolve(): │ │ ❱ 326 │ │ g_pipeline = create_pipeline( │ │ 327 │ │ │ base_model=base_model_path, │ │ 328 │ │ │ model_config=model_config, │ │ 329 │ │ │ infer_config=infer_config, │ │ │ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\src\animatediff\generate.py:371 in │ │ create_pipeline │ │ │ │ 368 │ │ logger.info(f"Loading weights from {model_path}") │ │ 369 │ │ if model_path.is_file(): │ │ 370 │ │ │ logger.debug("Loading from single checkpoint file") │ │ ❱ 371 │ │ │ unet_state_dict, tenc_state_dict, vae_state_dict = get_checkpoint_weights(mo │ │ 372 │ │ elif model_path.is_dir(): │ │ 373 │ │ │ logger.debug("Loading from Diffusers model directory") │ │ 374 │ │ │ temp_pipeline = StableDiffusionPipeline.from_pretrained(model_path) │ │ │ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\src\animatediff\utils\model.py:139 in │ │ get_checkpoint_weights │ │ │ │ 136 │ │ 137 def get_checkpoint_weights(checkpoint: Path): │ │ 138 │ temp_pipeline: StableDiffusionPipeline │ │ ❱ 139 │ temp_pipeline, _ = checkpoint_to_pipeline(checkpoint, save=False) │ │ 140 │ unet_state_dict = temp_pipeline.unet.state_dict() │ │ 141 │ tenc_state_dict = temp_pipeline.text_encoder.state_dict() │ │ 142 │ vae_state_dict = temp_pipeline.vae.state_dict() │ │ │ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\src\animatediff\utils\model.py:124 in │ │ checkpoint_to_pipeline │ │ │ │ 121 │ if target_dir is None: │ │ 122 │ │ target_dir = pipeline_dir.joinpath(checkpoint.stem) │ │ 123 │ │ │ ❱ 124 │ pipeline = StableDiffusionPipeline.from_single_file( │ │ 125 │ │ pretrained_model_link_or_path=str(checkpoint.absolute()), │ │ 126 │ │ local_files_only=True, │ │ 127 │ │ load_safety_checker=False, │ │ │ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\venv\lib\site-packages\diffusers\loaders.py:147 │ │ 1 in from_single_file │ │ │ │ 1468 │ │ │ │ force_download=force_download, │ │ 1469 │ │ │ ) │ │ 1470 │ │ │ │ ❱ 1471 │ │ pipe = download_from_original_stable_diffusion_ckpt( │ │ 1472 │ │ │ pretrained_model_link_or_path, │ │ 1473 │ │ │ pipeline_class=cls, │ │ 1474 │ │ │ model_type=model_type, │ │ │ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\venv\lib\site-packages\diffusers\pipelines\stab │ │ le_diffusion\convert_from_ckpt.py:1234 in download_from_original_stable_diffusion_ckpt │ │ │ │ 1231 │ │ │ # only refiner xl has embedder and one text embedders │ │ 1232 │ │ │ config_url = "https://raw.githubusercontent.com/Stability-AI/generative-mode │ │ 1233 │ │ │ │ ❱ 1234 │ │ original_config_file = BytesIO(requests.get(config_url).content) │ │ 1235 │ │ │ 1236 │ original_config = OmegaConf.load(original_config_file) │ │ 1237 │ │ │ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\venv\lib\site-packages\requests\api.py:73 in │ │ get │ │ │ │ 70 │ :rtype: requests.Response │ │ 71 │ """ │ │ 72 │ │ │ ❱ 73 │ return request("get", url, params=params, **kwargs) │ │ 74 │ │ 75 │ │ 76 def options(url, **kwargs): │ │ │ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\venv\lib\site-packages\requests\api.py:59 in │ │ request │ │ │ │ 56 │ # avoid leaving sockets open which can trigger a ResourceWarning in some │ │ 57 │ # cases, and look like a memory leak in others. │ │ 58 │ with sessions.Session() as session: │ │ ❱ 59 │ │ return session.request(method=method, url=url, **kwargs) │ │ 60 │ │ 61 │ │ 62 def get(url, params=None, **kwargs): │ │ │ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\venv\lib\site-packages\requests\sessions.py:587 │ │ in request │ │ │ │ 584 │ │ │ "allow_redirects": allow_redirects, │ │ 585 │ │ } │ │ 586 │ │ send_kwargs.update(settings) │ │ ❱ 587 │ │ resp = self.send(prep, **send_kwargs) │ │ 588 │ │ │ │ 589 │ │ return resp │ │ 590 │ │ │ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\venv\lib\site-packages\requests\sessions.py:701 │ │ in send │ │ │ │ 698 │ │ start = preferred_clock() │ │ 699 │ │ │ │ 700 │ │ # Send the request │ │ ❱ 701 │ │ r = adapter.send(request, **kwargs) │ │ 702 │ │ │ │ 703 │ │ # Total elapsed time of the request (approximately) │ │ 704 │ │ elapsed = preferred_clock() - start │ │ │ │ G:\A1111\Animediff\animatediff-cli-prompt-travel\venv\lib\site-packages\requests\adapters.py:553 │ │ in send │ │ │ │ 550 │ │ │ if isinstance(e.reason, ConnectTimeoutError): │ │ 551 │ │ │ │ # TODO: Remove this in 3.0.0: see #2811 │ │ 552 │ │ │ │ if not isinstance(e.reason, NewConnectionError): │ │ ❱ 553 │ │ │ │ │ raise ConnectTimeout(e, request=request) │ │ 554 │ │ │ │ │ 555 │ │ │ if isinstance(e.reason, ResponseError): │ │ 556 │ │ │ │ raise RetryError(e, request=request) │ ╰──────────────────────────────────────────────────────────────────────────────────────────────────╯ ConnectTimeout: HTTPSConnectionPool(host='raw.githubusercontent.com', port=443): Max retries exceeded with url: /CompVis/stable-diffusion/main/configs/stable-diffusion/v1-inference.yaml (Caused by ConnectTimeoutError(, 'Connection to raw.githubusercontent.com timed out. (connect timeout=None)')) (venv) G:\A1111\Animediff\animatediff-cli-prompt-travel>
-
Colab | Errors when installing x-formers
ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts. fastai 2.7.12 requires torch<2.1,>=1.7, but you have torch 2.1.0+cu118 which is incompatible. torchaudio 2.0.2+cu118 requires torch==2.0.1, but you have torch 2.1.0+cu118 which is incompatible. torchdata 0.6.1 requires torch==2.0.1, but you have torch 2.1.0+cu118 which is incompatible. torchtext 0.15.2 requires torch==2.0.1, but you have torch 2.1.0+cu118 which is incompatible. torchvision 0.15.2+cu118 requires torch==2.0.1, but you have torch 2.1.0+cu118 which is incompatible. WARNING[XFORMERS]: xFormers can't load C++/CUDA extensions. xFormers was built for: PyTorch 2.1.0+cu121 with CUDA 1201 (you have 2.1.0+cu118) Python 3.10.13 (you have 3.10.12) Please reinstall xformers (see https://github.com/facebookresearch/xformers#installing-xformers) Memory-efficient attention, SwiGLU, sparse and more won't be available. Set XFORMERS_MORE_DETAILS=1 for more details xformers version: 0.0.22.post3
-
FlashAttention-2, 2x faster than FlashAttention
This enables V1. V2 is still yet to be integrated into xformers. The team replied saying it should happen this week.
See the relevant Github issue here: https://github.com/facebookresearch/xformers/issues/795
-
Xformers issue
My Xformers doesnt work, any help see code. info ( Exception training model: 'Refer to https://github.com/facebookresearch/xformers for more information on how to install xformers'. ) or
-
Having xformer troubles
ModuleNotFoundError: Refer to https://github.com/facebookresearch/xformers for more
-
Question: these 4 crappy picture have been generated with the same seed and settings. Why they keep coming mildly different?
Xformers is a module that that can be used with Stable Diffusion. It decreases the memory required to generate an image as well as speeding things up. It works very well but there are two problems with Xformers:
-
Stuck trying to update xformers
WARNING[XFORMERS]: xFormers can't load C++/CUDA extensions. xFormers was built for: PyTorch 1.13.1+cu117 with CUDA 1107 (you have 2.0.1+cu118) Python 3.10.9 (you have 3.10.7) Please reinstall xformers (see https://github.com/facebookresearch/xformers#installing-xformers) Memory-efficient attention, SwiGLU, sparse and more won't be available. Set XFORMERS_MORE_DETAILS=1 for more details ================================================================================= You are running xformers 0.0.16rc425. The program is tested to work with xformers 0.0.17. To reinstall the desired version, run with commandline flag --reinstall-xformers. Use --skip-version-check commandline argument to disable this check. =================================================================================
-
Question about updating Xformers for A1111
# Your version of xformers is 0.0.16rc425. # xformers >= 0.0.17.dev is required to be available on the Dreambooth tab. # Torch 1 wheels of xformers >= 0.0.17.dev are no longer available on PyPI, # but you can manually download them by going to: https://github.com/facebookresearch/xformers/actions # Click on the most recent action tagged with a release (middle column). # Select a download based on your environment. # Unzip your download # Activate your venv and install the wheel: (from A1111 project root) cd venv/Scripts activate pip install {REPLACE WITH PATH TO YOUR UNZIPPED .whl file} # Then restart your project.
-
Is there a Pygmalion roadmap?
Further reading/resources: RedPajama: https://www.together.xyz/blog/redpajama xFormers: https://github.com/facebookresearch/xformers Flash Attention: https://arxiv.org/abs/2205.14135 Sparsity [NEW!]: https://arxiv.org/abs/2304.07613
- Slow/short replies?
pytorch-image-models
- FLaNK AI Weekly 18 March 2024
-
[D] Hugging face and Timm
I am a PyTorch user I work in CV, I usually use the PyTorch models. However, I see people use timm in research papers to train their models I don't understand what it is timm is it a new framework like PyTorch? Further, when I click https://pypi.org/project/timm/ homepage it takes me to hugging face GitHub https://github.com/huggingface/pytorch-image-models is there any connection between timm and hugging face many of my friends use hugging face but I also don't know about hugging face I use simple PyTorch and torchvision.models.
-
FLaNK Stack Weekly for 07August2023
https://github.com/huggingface/pytorch-image-models https://huggingface.co/docs/timm/index
-
[R] Nvidia RTX 4090 ML benchmarks. Under QEMU/KVM. Image + Transformers. FP16/FP32.
pytorch-image-models
-
Inference on resent, cant work out the problem?
additionally, you might find the timm library handy for this sort of work.
-
Swin Transformer: Hierarchical Vision Transformer Using Shifted Windows
This is still being pursued. Ross Wightmann's timm[0,1] package (now on Hugging Face) has done a lot of this. There's also a V2 of ConvNext[2]. Ross does write about this a lot on Twitter fwiw. I should also mention that there are still many transformer based networks that still beat convs. So there probably won't be a resurgence in convs until someone can show that there's a really strong reason for them. They have some advantages but they also might not be flexible enough for the long range tasks in segmentation and detection. But maybe they are.
FAIR definitely did great work with ConvNext, and I do hope to see more. There always needs to be people pushing unpopular paradigms.
[0] https://github.com/huggingface/pytorch-image-models
[1] https://arxiv.org/abs/2110.00476
[2] https://arxiv.org/abs/2301.00808
-
Problems with Learning Rate Finder in Pytorch Lightning
I am doing Binary classification with a pre-trained EfficientNet tf_efficientnet_l2. I froze all weights during training and replaced the classifier with a custom trainable one that looks like:
-
PyTorch at the Edge: Deploying Over 964 TIMM Models on Android with TorchScript and Flutter
In this post, I’m going to show you how you can pick from over 900+ SOTA models on TIMM, train them using best practices with Fastai, and deploy them on Android using Flutter.
-
ImageNet Advise
The other thing is, try to find tricks to speed up your experiments (if not having done so already). The most obvious are mixed precision training, have your model train on a lower resolution input first and then increase the resolution later in the training, stochastic depth, and a bunch more stuffs. Look for implementations in https://github.com/rwightman/pytorch-image-models .
- Doubt about transformers
What are some alternatives?
flash-attention - Fast and memory-efficient exact attention
yolov5 - YOLOv5 🚀 in PyTorch > ONNX > CoreML > TFLite
stable-diffusion-webui - Stable Diffusion web UI
mmdetection - OpenMMLab Detection Toolbox and Benchmark
SHARK - SHARK - High Performance Machine Learning Distribution
detectron2 - Detectron2 is a platform for object detection, segmentation and other visual recognition tasks.
Dreambooth-Stable-Diffusion - Implementation of Dreambooth (https://arxiv.org/abs/2208.12242) with Stable Diffusion
mmcv - OpenMMLab Computer Vision Foundation
diffusers - 🤗 Diffusers: State-of-the-art diffusion models for image and audio generation in PyTorch
segmentation_models.pytorch - Segmentation models with pretrained backbones. PyTorch.
InvokeAI - InvokeAI is a leading creative engine for Stable Diffusion models, empowering professionals, artists, and enthusiasts to generate and create visual media using the latest AI-driven technologies. The solution offers an industry leading WebUI, supports terminal use through a CLI, and serves as the foundation for multiple commercial products.
yolact - A simple, fully convolutional model for real-time instance segmentation.