Unable to find candidates for torch #10125
Description
Description
So I am trying to install torch on one of my projects and my specs are as below
Python 3.12.3 / pip latest / Poetry (version 2.0.1)/ I am on Apple M3 Max
Here is my pyproject.toml
torch = { version = "^2.5.1+cpu", source = "pytorch-cpu" }
torchvision = { version = "^0.20.1+cpu", source = "pytorch-cpu" }
torchaudio = { version = "^2.5.1+cpu", source = "pytorch-cpu" }
[[tool.poetry.source]]
name = "pytorch-cpu"
url = "https://download.pytorch.org/whl/cpu"
priority = "explicit"
Adding a screenshot for more context
When I create a new project I am able to install with pip torch up to 2.6.0(which is latest)
below will be my full dependency list just in case
[tool.poetry.dependencies]
python = "~3.12.3"
fastapi = "^0.111.0"
uvicorn = "^0.30.1"
loguru = "^0.7.2"
sqlalchemy = "^2.0.30"
boto3 = "^1.34.122"
asyncpg = "^0.29.0"
greenlet = "^3.0.3"
python-dotenv = "^1.0.1"
jwcrypto = "^1.5.6"
jinja2 = "^3.1.4"
mangum = "^0.17.0"
requests = "^2.32.3"
deepgram-sdk = "^3.2.7"
langchain-openai = "^0.3.1"
langchain-google-genai = "^2.0.9"
langchain-mistralai = "^0.2.4"
langchain-anthropic = "^0.3.3"
google-auth-oauthlib = "^1.2.0"
fuzzywuzzy = "^0.18.0"
levenshtein = "^0.25.1"
msgraph-sdk = "^1.4.0"
twilio = "^9.1.1"
alembic = "^1.13.1"
python-multipart = "^0.0.9"
langchain = "^0.3.15"
langgraph = "^0.2.66"
pytz = "^2024.1"
bcrypt = "^4.1.3"
langchain-core = "^0.3.31"
langchain-text-splitters = "^0.3.5"
langsmith = "^0.3.1"
langchain-community = "^0.3.15"
langchain-cohere = "^0.4.1"
# langchain-google-vertexai = "^1.0.5"
langchain-postgres = "^0.0.12"
s3fs = "^2024.6.1"
azure-storage-blob = "^12.20.0"
sqlparse = "^0.5.0"
pymupdf = "^1.24.7"
ultralytics = "^8.2.60"
opencv-python-headless = "^4.10.0.84"
llm-sdk = "^0.3.2"
sentry-sdk = { extras = ["fastapi"], version = "^2.13.0" }
nltk = "^3.9.1"
unstructured = { extras = ["all-docs"], version = "^0.16.11" }
unstructured-ingest = { extras = ["gcs"], version = "^0.1.1" }
matplotlib = "^3.9.2"
python-magic = "^0.4.27"
pytesseract = "^0.3.13"
psycopg = { extras = ["binary"], version = "^3.2.2" }
pydantic = "2.10.5"
torch = { version = "^2.5.1+cpu", source = "pytorch-cpu" }
torchvision = { version = "^0.20.1+cpu", source = "pytorch-cpu" }
torchaudio = { version = "^2.5.1+cpu", source = "pytorch-cpu" }
[[tool.poetry.source]]
name = "pytorch-cpu"
url = "https://download.pytorch.org/whl/cpu"
priority = "explicit"
[build-system]
requires = ["poetry-core"]
build-backend = "poetry.core.masonry.api"
Workarounds
No idea
Poetry Installation Method
pipx
Operating System
MacOS 15.2
Poetry Version
2.0.1
Poetry Configuration
cache-dir = "/Users/userName/Library/Caches/pypoetry"
installer.max-workers = null
installer.no-binary = null
installer.only-binary = null
installer.parallel = true
installer.re-resolve = true
keyring.enabled = true
repositories.pytorch-cpu.url = "https://download.pytorch.org/whl/cpu"
requests.max-retries = 0
solver.lazy-wheel = true
system-git-client = false
virtualenvs.create = true
virtualenvs.in-project = null
virtualenvs.options.always-copy = false
virtualenvs.options.no-pip = false
virtualenvs.options.system-site-packages = false
virtualenvs.path = "{cache-dir}/virtualenvs" # /Users/UserName/Library/Caches/pypoetry/virtualenvs
virtualenvs.prompt = "{project_name}-py{python_version}"
virtualenvs.use-poetry-python = falsePython Sysconfig
sysconfig.log
Paste the output of 'python -m sysconfig', over this line.
Example pyproject.toml
Poetry Runtime Logs
poetry-runtime.log
- Installing torch (2.5.1+cpu): Failed
RuntimeError
Unable to find installation candidates for torch (2.5.1+cpu)
at ~/.local/pipx/venvs/poetry/lib/python3.12/site-packages/poetry/installation/chooser.py:86 in choose_for
82│
83│ links.append(link)
84│
85│ if not links:
→ 86│ raise RuntimeError(f"Unable to find installation candidates for {package}")
87│
88│ # Get the best link
89│ chosen = max(links, key=lambda link: self._sort_key(package, link))
90│
Cannot install torch.