Model Training & Fine-tuning
- orchestra-research BundleMoe Training
Train Mixture of Experts (MoE) models using DeepSpeed or HuggingFace, covering architectures, routing, load balancing, and expert parallelism.
14.5k - orchestra-research BundleSpeculative Decoding
Accelerate LLM inference using speculative decoding, Medusa multiple heads, and lookahead decoding techniques for 1.5-3.6× speedup without quality loss.
14.5k - qhjqhj00 BundleHf MCP
Connects AI assistants to the Hugging Face Hub via MCP server tools to search models, datasets, Spaces, and papers, retrieve repository details and documentation, run compute jobs, and use Gradio Spaces as AI tools.
3 - k-dense-ai BundlePymc
Build, fit, validate, and compare Bayesian models using PyMC's modern API, including hierarchical models, MCMC sampling, variational inference, posterior predictive checks, and model comparison.
57.3k - k-dense-ai BundleGeniml
Train unsupervised machine learning models on genomic interval data from BED files, including region embeddings, single-cell ATAC-seq analysis, and consensus peak building.
57.3k - orchestra-research BundleModel Pruning
Compress large language models by 40-60% with minimal accuracy loss using one-shot pruning techniques like Wanda and SparseGPT, enabling faster inference and deployment on constrained hardware.
14.5k - dvcrn BundleSoma
Guides users through participating in the SOMA decentralized training network, covering data submission, model training, reward claiming, and strategic optimization.
33 - clawbio BundleGi Enhancer
Predict enhancer activity in DNA sequences using the Genomic Intelligence G0 DeepSTARR model, via the hosted /v1/tasks/enhancer/predict API. Returns per-window activity scores.
1.2k - clawbio BundleGi Chromatin
Predict chromatin state — histone marks, DNase, TF binding — across 919 tracks (DeepSEA-style) for DNA sequences, via the hosted Genomic Intelligence /v1/tasks/chromatin/predict API.
1.2k - clawbio BundleGi Annotation
Predict gene and transcript structure (intervals, exons, strand) from a DNA sequence using the Genomic Intelligence DNA Annotation model, via the hosted /v1/tasks/annotation/predict API. Submitted asynchronously — the pipeline takes ~20 s for ~20 kbp.
1.2k - orchestra-research BundleDeepspeed
Provides expert guidance for distributed training with DeepSpeed, covering ZeRO optimization stages, pipeline parallelism, FP16/BF16/FP8, 1-bit Adam, and sparse attention.
14.5k - google-gemma BundleGemma Dev
Selects the right Gemma model for a task, recommends deployment tooling (Gradio, Transformers.js, Vertex AI, MLX), and applies optimizations like MTP and QAT.
- google-gemma BundleGemma Trainer
Fine-tune Gemma models locally using QLoRA, Unsloth, or TRL for SFT, DPO, and reward modeling, with dataset preparation and conversion to GGUF or LiteRT-LM.
- clawbio BundleGi Expression
Predict tissue / cell-type expression (log TPM + TPM) from 9,198–500,000 bp of DNA around a TSS, at least 4,599 bp each side (anything but exactly 9,198 bp needs --tss-index) using the Genomic Intelligence G0 Expression model, via the hosted /v1/tasks/expression/predict API. The model is conditioned on a free-text cell-type / assay description.
1.2k - orchestra-research SkillLlamaguard
Deploy Meta's LlamaGuard moderation model to filter LLM inputs and outputs across 6 safety categories using HuggingFace, vLLM, or FastAPI.
14.5k