Tag
Explore by tags
Provides ~5M model-generated reasoning chains (within 5k sequence length) with structured fields for supervised fine-tuning, reasoning distillation, and instruction tuning. Includes separate fields for prompt, reasoning trace, final answer and a ChatML view; streaming access recommended for large-scale use.
Provides ~5M tokens of chain-of-thought reasoning traces generated by many LLMs (DeepSeek, Qwen, Gemma, etc.) for training and evaluating reasoning SLMs — includes repo_id, tok_len, user, thought_trace, assistant and ChatML fields; sequences limited to 5k.
Performs full-parameter post-training of trillion-parameter MoE DeepSeek-V4 models on an Ascend NPU SuperPOD, using a hierarchical optimization of model parallelism, communication orchestration, and kernel execution to increase Model FLOPs Utilization. Also builds CPT/SFT pipelines with solver-verified synthetic data for Operations Research, reporting strong zero-shot Pass@1 results.
An OpenAI-compatible LLM checkpoint optimized for agentic and long-context scenarios, shipping DSpark speculative decoding and vLLM/SGLang deployment recipes; tailored for code-agent and multi-step reasoning workloads and released under MIT.
Enables local use of a GGUF-quantized DeepSeek-V4-Flash-0731 via Unsloth Dynamic quantizations; provides a Q8 (162GB) lossless option and smaller Q4 variants for lower-memory inference and agentic scenarios using Unsloth tooling.
Provides a Mixture-of-Experts language model tuned for million-token contexts and agentic workflows, with DSpark speculative decoding, FP4/FP8 mixed-precision support, and vLLM/SGLang deployment recipes for low-latency production inference.
A synthetic, verifiable-first agentic training corpus with 19,072 training traces and 2,135 held-out evaluation rows. Provides per-turn visible reasoning, real sandboxed tool executions, 13 verifiable task families, and NeMo Gym / RL-ready reward contracts for SFT and RL workflows.
Enables interactive serving of large Mixture-of-Experts (MoE) models on personal machines by adapting offload and execution to measured device bandwidth and agentic workload patterns. Key features include bandwidth-adaptive execution, semantic-aware caching of recurrent state, and an elastic GPU expert cache; supports 20+ MoE models and runs models from ~35B to 753B on consumer/workstation GPUs.
Synthesizes, repairs, and self-evolves task-adaptive agent harnesses on demand for off-the-shelf LLM agents, using a trainable harness-intelligence model that distills signals from past configurations. Demonstrates consistent performance gains across benchmarks and model families by producing four-module, composable harnesses.
An experimental multimodal model that adds visual understanding to DeepSeek-V4-Flash: accepts text+image inputs and returns text analyses. Improves vision-dependent agent workflows while maintaining comparable text-only performance; released under an MIT license on Hugging Face.
Turns each research paper into a training environment to generate verifiable research plans by synthesizing questions from goals/background and deriving evaluation criteria from methods/experiments. Key features: four-stage extraction that reduces criterion leakage to 3.7%, a two-stage rubric-centered training (self-distillation then GRPO), and the PaperGym-20k corpus with two held-out benchmarks.
A multimodal Mixture-of-Experts foundation model with a million-token context window; uses a causal encoder–decoder layout and aggressive KV-cache compression (~890 bytes/token) to limit per-token activation to 8B/16B—designed for long-context, agentic, and multimodal workloads.