Packs
1 packResults for “model-training”
21 skills
jiachen-t-wang training-compute-optimal-large-language-models-arxiv-2203-15
Training Compute-Optimal Large Language Models
6
jiachen-t-wang deduplicating-training-data-makes-language-models-better-arx
Deduplicating Training Data Makes Language Models Better
6
jiachen-t-wang vila-on-pre-training-for-visual-language-models-arxiv-2312-0
VILA: On Pre-training for Visual Language Models
6
paramchordiya ml-engineering
Enforces rigorous ML modeling, feature engineering, training, and evaluation standards at principal-engineer level.
0
smith6jt-cop crypto-database-population
Populate symbol database with crypto symbols before training. Trigger when: (1) live trader fails with 'CRYPTO VIOLATION', (2) no crypto models in models/rl_symbols/, (3) db.get_candidates(asset_types=['crypto']) returns 0, (4) starting fresh training without crypto.
3
tianhao909 sparse-autoencoder-training
Provides guidance for training and analyzing Sparse Autoencoders (SAEs) using SAELens to decompose neural network activations into interpretable features. Use when discovering interpretable features, analyzing superposition, or studying monosemantic representations in language models.
1 · bundle
More results
qcmuu sparse-autoencoder-training
Provides guidance for training and analyzing Sparse Autoencoders (SAEs) using SAELens to decompose neural network activations into interpretable features. Use when discovering interpretable features, analyzing superposition, or studying monosemantic representations in language models.
0 · bundle
jiachen-t-wang scaling-data-constrained-language-models-arxiv-2305-16264v3
Scaling Data-Constrained Language Models
6
jiachen-t-wang chameleon-mixed-modal-early-fusion-foundation-models-arxiv-2
Chameleon: Mixed-Modal Early-Fusion Foundation Models
6
jiachen-t-wang emu2-generative-multimodal-models-are-in-context-learners-ar
Emu2: Generative Multimodal Models are In-Context Learners
6
jiachen-t-wang matryoshka-representation-learning-arxiv-2205-13147v4
Matryoshka Representation Learning
6
jiachen-t-wang eva-clip-improved-training-techniques-for-clip-at-scale-arxi
EVA-CLIP: Improved Training Techniques for CLIP at Scale
6
jiachen-t-wang influence-functions-in-deep-learning-arxiv-2002-08484v3
Influence Functions in Deep Learning
6
jiachen-t-wang scaling-instruction-finetuned-language-models-arxiv-2210-114
Scaling Instruction-Finetuned Language Models
6
tianhao909 openrlhf-training
High-performance RLHF framework with Ray+vLLM acceleration. Use for PPO, GRPO, RLOO, DPO training of large models (7B-70B+). Built on Ray, vLLM, ZeRO-3. 2× faster than DeepSpeedChat with distributed architecture and GPU resource sharing.
1 · bundle
qcmuu openrlhf-training
High-performance RLHF framework with Ray+vLLM acceleration. Use for PPO, GRPO, RLOO, DPO training of large models (7B-70B+). Built on Ray, vLLM, ZeRO-3. 2× faster than DeepSpeedChat with distributed architecture and GPU resource sharing.
0 · bundle
lord1egypt llava
Runs the open-source LLaVA vision-language model for image understanding, captioning, visual question answering, and multi-turn image conversations, including setup, inference, and training guidance.
2
orchestra-research pytorch-fsdp2
Adds PyTorch FSDP2 (fully_shard) to training scripts with correct init, sharding, mixed precision/offload config, and distributed checkpointing. Use when models exceed single-GPU memory or when you need DTensor-based sharding with DeviceMesh.
10.4k · bundle
orchestra-research openrlhf-training
Train large language models (7B-70B+) with RLHF using PPO, GRPO, DPO, and other algorithms, accelerated by Ray and vLLM for distributed multi-GPU setups.
10.4k · bundle
tianhao909 pytorch-fsdp2
Adds PyTorch FSDP2 (fully_shard) to training scripts with correct init, sharding, mixed precision/offload config, and distributed checkpointing. Use when models exceed single-GPU memory or when you need DTensor-based sharding with DeviceMesh.
1 · bundle
qcmuu pytorch-fsdp2
Adds PyTorch FSDP2 (fully_shard) to training scripts with correct init, sharding, mixed precision/offload config, and distributed checkpointing. Use when models exceed single-GPU memory or when you need DTensor-based sharding with DeviceMesh.
0 · bundle