Tag
Explore by tags
Runs an autonomous self-improvement loop where a meta agent crafts a task-specific agent, a target agent executes trials, and a feedback agent updates both harness (code) and model weights—provider-agnostic profiles with reproducible runs and a live dashboard.
Provides 1,000,000 model-generated chain-of-thought traces and instruction–response pairs for fine-tuning and distilled supervision. Focused splits (coding, PHD-Science, General-Math, MultilingualSTEM), ~5B tokens, Apache-2.0 license.
Provides a unified 615k-hour English speech corpus for TTS training, aggregating 11 public datasets and web-sourced recordings into 16 kHz Opus WebDataset shards. Includes a quality-filtered core subset (510.1k hours), metadata splits, and mixed licenses across sources.
Aggregates and deduplicates public Claude distillation datasets into a unified 'messages' format with source attribution; focused on instruction-tuning and reasoning samples for SFT and LLM training, while requiring users to follow original sources' licenses.
Provides 12.26M synthetically generated multilingual OCR samples (en/ja/ko/ru/zh) with word/line/paragraph bounding boxes and reading-order graphs, packaged as HDF5 shards for training detection, recognition, and layout models; licensed CC BY 4.0.
Provides a ~9.2M-instance Japanese multimodal post-training dataset for vision–language models, combining image–text pairs, PDF corpora and generated VQA to improve Japanese VLM performance; access is restricted by Japanese copyright (download via llm-jp GitLab).
Generates text by iteratively denoising blocks of tokens with a two-tower design: a frozen autoregressive context tower and a trainable diffusion denoiser tower, trading minimal quality loss for higher wall-clock throughput.
A Mixture-of-Experts instruct-capable LLM (295B total, 21B active) designed for long-context reasoning, code/agent workflows and instruction-following; released by Tencent Hy Team with safetensors weights on Hugging Face.
Provides deduplicated, sanitized Usenet posts (1980–2013) for language-model pretraining and linguistic research. Includes a ~103.1B-token full corpus (408M posts) with freely downloadable sample files; full corpus access requires a license and PII redaction was applied.
Provides ~12.29M execution‑free agentic coding trajectories (≈112B tokens) sampled from 122K GitHub PRs to mid‑train code and agent models. Uses bash-only actions (grep, git, sed, etc.) so it scales without Docker; trajectories are unverified and intended for mid-training rather than final SFT.
Provides 207k+ LLM-generated agent trajectories of code edits and tool interactions for training and evaluating software-engineering agents. Collected via OpenHands and SWE-agent using Qwen3.5-122B and MiniMax-M2.5, multilingual across nine languages and released under CC BY 4.0.
A healed 64-layer 'frankenmerge' that stacks two Qwen3.5-derived finetunes into an ~18B GGUF model for multilingual text generation, reasoning, and reliable code/frontend output. Healed with a 1000-step QLoRA to reduce layer-boundary artifacts and targeted to run on 12–16 GB GPUs.