smcleish/Recurrent-Llama-3.2-2-4-2-untrained
smcleish/Recurrent-Llama-3.2-2-4-2-untrained is an open-weights language model for text generation under the apache-2.0 license (4.1 GB), added to Hugging Bay Jun 30, 2026. Hosted open-weights model — 6 files carry published SHA-256 checksums you can re-hash.
A chat model — but here it’s a set of files: download them and load them with a Python library such as Hugging Face transformers on a computer.
About this download
Recurrent-Llama-3.2-2-4-2-untrained. 1 file. Published by smcleish
How this works
To run a downloaded model you need the file and a runner program to open it. Searching and reading need no account.
You don’t need to know what a “model” is. This is the one for chatting.
Not a fact-checker. Not a search engine. It can be confidently wrong.
Not sure what you’d use it for? Try one question — that’s the whole test. Try: “Explain photosynthesis like I’m five.”
Downloadable — the files passed our public gates and the download button is here today. That’s all it means — not safe, not guaranteed to run.
Checked on this Mac — the file’s fingerprint matches the published one, byte for byte. That proves it arrived unchanged — not that the model is safe.
Running locally — this device did the work: a real first reply was produced here — not on a server.
Keep the file
Download · 4.08 GB (1 file) saves to your Downloads folder
Run on a computer
On this page you get the files to download, a plain explanation of the model, and the checks that prove the files arrived unchanged.
To run a downloaded model you need the file and a runner app to open it.
What happens next
- Download the file (4.08 GB).
- To run a downloaded model you need the file and a runner program to open it.
- Then type a question.
model.safetensors · 4.08 GB
DETAILS
Checked. Ready to download.
Make sure it arrived whole. Checking a 4.1 GB file can take a few minutes
If the string it prints matches this one, you're holding exactly what we published:
269a...1acb
Think of the fingerprint below as the file's ID photo: a string unique to its exact bytes.
What this step does: it only prints an ID string so you can compare; it doesn't run anything or install anything.
On a Mac, the Terminal line below prints this file's SHA-256 fingerprint; it already contains your file's name and assumes the file is in your Downloads folder.
After the download finishes, paste it into Terminal and press Return.
Terminal shasum -a 256 ~/Downloads/model.safetensors
On a different computer?
On Windows, the PowerShell line below prints this file's SHA-256 fingerprint; it already contains your file's name and assumes the file is in your Downloads folder.
After the download finishes, paste it into PowerShell and press Enter.
PowerShell Get-FileHash -Algorithm SHA256 "$HOME\Downloads\model.safetensors"
From Command Prompt instead: certutil -hashfile "%USERPROFILE%\Downloads\model.safetensors" SHA256
PowerShell prints the same SHA-256 fingerprint in capital letters.
On Linux, the terminal line below prints this file's SHA-256 fingerprint; it already contains your file's name and assumes the file is in your Downloads folder.
After the download finishes, paste it into a terminal and press Enter.
Terminal sha256sum ~/Downloads/model.safetensors
File fingerprint (for experts)
269ac86c4c6d35b2c181ce29032b26c1176306a90595e4e3ac982508b85c1acb
This check proves the file arrived unchanged. Not that the model is safe.
What we've checked:
- License. Passed (apache-2.0)
- Fingerprint checked
- Files hosted. 6 of 10 from the repository
What does ‘Checked’ mean?
Checked on this Mac — the file’s fingerprint matches the published one, byte for byte. That proves it arrived unchanged — not that the model is safe.
Other files (for experts)
- .gitattributes · 2 KB
- README.md · 2 KB
- config.json · 2 KB
- generation_config.json · 144 bytes
- raven_config_minimal.py · 4 KB
- raven_modeling_minimal.py · 74 KB
- raven_modeling_minimal_with_init.py · 78 KB
- special_tokens_map.json · 301 bytes
- tokenizer_config.json · 51 KB
- Task
- text generation
- Recommended file; this is what downloads
- 4.1 GB — model.safetensors
- Size
- Recorded total: 4.1 GB · selected file: 4.1 GB
- Files
- 6 hosted of 10 listed
- Download
- Download here (6 hosted files)
- Added to Hugging Bay
- Updated at source
FAQ
What is this model?
smcleish/Recurrent-Llama-3.2-2-4-2-untrained is an open-weights language model for text generation published by smcleish, added to Hugging Bay Jun 30, 2026. Model from Hugging Face for text generation: smcleish/Recurrent-Llama-3.2-2-4-2-untrained · transformers
Can I use it commercially?
It is released under the apache-2.0 license. Commercial use is allowed. Permissive open license: commercial use, modification, and redistribution are permitted.
How big is it?
The total download is 4.1 GB; the recommended download, model.safetensors, is 4.1 GB.
What can I download here?
Hugging Bay hosts 6 of 10 files for this model, each with a published SHA-256 checksum. Hosted download size: 4.1 GB (full repository: 4.1 GB). The original release is at https://huggingface.co/smcleish/Recurrent-Llama-3.2-2-4-2-untrained.
File checks and computer fit
File pattern scans
Automated file-pattern scans flagged nothing — not a safety guarantee. 6 hosted files carry a published SHA-256 you can re-hash. The machine-readable verdict is at /api/artifacts/hf-model-smcleish-recurrent-llama-3-2-2-4-2-untrained/safety on Hugging Bay.
Will smcleish/Recurrent-Llama-3.2-2-4-2-untrained run on my computer?
We can’t check the fit — we don’t know your machine. We don’t have your machine’s specs, so no check ran and no verdict is shown. Answer two quick questions and we’ll estimate, from the memory you tell us, whether it fits your machine. Or skip it — you can still download; fit isn’t checked.
About this artifact
External Hugging Face catalog record for smcleish/Recurrent-Llama-3.2-2-4-2-untrained. Task: text generation. Library: transformers. Languages: en. Popularity signal at import: 15 downloads and 1 likes.
Hosted file inventory
- Hosted files
- 6 of 10
- Reviewed peer-assisted fallbacks
- 0
- Trust bundle
- Trust-bundle JSON
- Review summary
- Review-summary JSON
- Distribution
- Reviewed peer-fallback JSON
Run status
Detailed evidence for the selected file set
This detail applies to the selected file set, not every file in the repository.
- Executable state
- unknown
- Evidence timestamp
- unknown
- Selected representation
- unknown
- Complete bytes
- unknown
- Complete kit
- unknown
- Byte-integrity check
- unknown
- Run ready
- unknown
- Files
- View file metadata
Bay Run
Bay Run pins; Try It Now unavailable
Bay Run does not publish a serving contract for this exact artifact and Hugging Bay will not invent one.
Mint a Pin unavailable
Bay Run does not publish a serving contract for this exact artifact and Hugging Bay will not invent one.
Guard before generate
Add Bay Run's provider-neutral coprocessor to your own generation pipeline. It checks decision.action before your generator runs.
npm install @huggingbay/coprocessor · Install @huggingbay/coprocessor
This handoff wraps your caller-owned generator. It does not run this catalog artifact, make it a Bay Run Pin, or change its serving or verification state.
Request a complete representation
- License: pass - apache-2.0
- Provenance: pass - smcleish/Recurrent-Llama-3.2-2-4-2-untrained
- Scan: pass - passed
- Hosted files: warn - 6/10 repository files hosted; selected representation evaluated separately
- Byte manifest integrity: warn - not final yet
Machine-readable card
Open interactive artifact page
Evidence details
- Demand
- 22 upstream downloads
- Trust
- Unknown
- Run fit
- Unknown: canonical-state-temporarily-unavailable
- Scan status
- passed
- Executable state
- unknown
- Evidence timestamp
- unknown
- Declared license
- apache-2.0
- Pipeline
- text-generation
- Library
- transformers
- Task categories
- text-generation
- Languages
- en
- Paper ids
- 2511.07384
- File hints
- 11 upstream file samples