research-artifacts
dlm-exp1-drafter-voting-research-artifacts
dlm-exp1 drafter-voting: research artifacts (public mirror)
Everything bulky from the drafter-voting branch of https://github.com/NoviceCoderInfinity/dlm-general-expert-exp1
that git does not track (uploaded 2026-09-06 before the compute instance was destroyed):
logs/research_v1/: baseline cells (one JSON line per item; per-pass NPZ traces for both checkpoints in
baselines_traces/ and baselines_b32_traces/), state-batch fidelity probe states (fidelity/), flush-fusion
boundaries… See the full description on the dataset page: https://huggingface.co/datasets/Arushhh/dlm-exp1-drafter-voting-research-artifacts.dlm-exp1-drafter-voting-research-artifacts
dlm-exp1 drafter-voting research artifacts (public)
Draft-aligned judge LoRA for SDAR-30B-A3B-Chat-b32 with a 1.7B drafter and DES expert pooling: goal = decode speedup with minimal accuracy loss vs the stock model.
model / setting
weights
eval records
stock SDAR-30B-A3B-Chat-b32 (+ 1.7B drafter where used)
JetLM/SDAR-30B-A3B-Chat-b32 @ c351bbc3… (Hugging Face)
reference_evals/da_decode (*_stock), r3_A/evals/stock
stock + DES32 / DES48 (decode-time expert pooling… See the full description on the dataset page: https://huggingface.co/datasets/Anupam-Rawat-IITB/dlm-exp1-drafter-voting-research-artifacts.otu-taxa-paper-artifacts
OTU-Taxa paper artifacts
This repository contains stable processed artifacts used by the OTU-Taxa paper
and the frozen contracts needed to reproduce them.
Related repositories
Source code:
https://github.com/bio-ontology-research-group/otu-taxa-autoregressive
Comparator pipelines:
https://github.com/bio-ontology-research-group/microbiome_foundation_model_benchmarks
Pretrained OTU-Taxa model:
bio-ontology-research-group/otu-taxa
Model weights and reusable… See the full description on the dataset page: https://huggingface.co/datasets/bio-ontology-research-group/otu-taxa-paper-artifacts.reasoning-switch-research-artifactsresearch-artifacts-archive
Research Artifacts Archive
This dataset contains an encrypted archive of code, experiment metrics, configuration files, notebooks, and explicitly selected model/checkpoint directories.
The payload is encrypted with GnuPG symmetric AES-256 encryption and split into ordered shards for reliable Hub storage. Hugging Face cannot preview the encrypted contents.
Files
data/data-00000-of-00022.tar.zst.gpg through data/data-00021-of-00022.tar.zst.gpg
SHA256SUMS contains a… See the full description on the dataset page: https://huggingface.co/datasets/vatolinalex/research-artifacts-archive.GLM-5.3-Flash-Research-Artifacts
GLM-5.3 Flash research artifacts
This is a research dataset from Alexey's attempt to make GLM-5.3 Flash useful on 4x RTX 3090. It contains numerical measurements, synthetic reference tests and a license-reviewed subset of teacher distributions captured on RTX 6000 PRO hardware. The experiment did not produce a quality-qualified production model.
The teacher was RedHatAI/GLM-5.3-Flash-NVFP4, revision 36c184c6cda000a481711306df5adde42f63321a, based on Z.AI's GLM-5.3 Flash. These… See the full description on the dataset page: https://huggingface.co/datasets/anonymousmaharaj/GLM-5.3-Flash-Research-Artifacts.