Llama historical research checkpoints

Built with Llama.

This repository preserves completed local research model and adapter artifacts. Each checkpoints/<id>/ directory is one independent checkpoint; it is not a standalone combined model. File hashes and original artifact names are recorded in manifests/<id>.json.

The original checkpoint files, configuration, tokenizer, and bundled custom modeling code are preserved byte-for-byte. These are research variants trained or modified from Llama 3.1. Consult each original configuration for the exact base checkpoint: some adapters require a previously trained research base rather than the initial Meta model. Each checkpoint retains its original documentation and applicable third-party notices. The Llama 3.1 license and acceptable-use policy apply to the Llama-derived materials. Existing checkpoint-specific license notices are also retained.

No new quality evaluation or training-resume validation is claimed by this archive. Some artifacts contain only adapters or partial model states; consult their configuration and original loader. Use a pinned commit and verify every recorded SHA-256 before restoring files.

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for jeongseokoh/Llama-SPEED-Capsule-archive-20261024

Finetuned
(3293)
this model

Collection including jeongseokoh/Llama-SPEED-Capsule-archive-20261024