Hugging Face's logo Hugging Face
  • Models
  • Datasets
  • Spaces
  • Buckets new
  • Docs
  • Enterprise
  • Pricing
    • Website
      • Tasks
      • HuggingChat
      • Collections
      • Languages
      • Organizations
    • Community
      • Blog
      • Posts
      • Daily Papers
      • Learn
      • Discord
      • Forum
      • GitHub
    • Solutions
      • Team & Enterprise
      • Hugging Face PRO
      • Enterprise Support
      • Inference Providers
      • Inference Endpoints
      • Storage Buckets

  • Log In
  • Sign Up

Codeseys
/
composer-replication-framework

Reinforcement Learning
Transformers
English
post-training
distillation
agentic-coding
composer-2.5
cursor
kimi-k2
grpo
dapo
diloco
openenv
trl
verl
research
methodology
Model card Files Files and versions
xet
Community

Instructions to use Codeseys/composer-replication-framework with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

  • Libraries
  • Transformers

    How to use Codeseys/composer-replication-framework with Transformers:

    # Load model directly
    from transformers import AutoModel
    model = AutoModel.from_pretrained("Codeseys/composer-replication-framework", dtype="auto")
  • Notebooks
  • Google Colab
  • Kaggle
composer-replication-framework / spikes
385 kB
Ctrl+K
Ctrl+K
  • 2 contributors
History: 9 commits
Codeseys's picture
Codeseys
Wave 18: 14 backlog items closed + 3-reviewer cross-family review
54efac8 1 day ago
  • 001-teacher-replay-cost
    Wave 18: 14 backlog items closed + 3-reviewer cross-family review 1 day ago
  • 002a-mini-gpu-smoke
    Wave 12: close V1-V8 brief โ€” GPU smoke, SDPO firing, real-trace e2e 2 days ago
  • 002a-trace-collection-trl
    Spike v0.0 laydown + spike 001 VALIDATED 2 days ago
  • 002b-trace-collection-prime-rl
    Spike v0.0 laydown + spike 001 VALIDATED 2 days ago
  • 003-dpo-pairs-from-disagreement
    Spike v0.0 laydown + spike 001 VALIDATED 2 days ago
  • 004-ab-train-grpo-vs-trace-replay-dpo
    Spike v0.0 laydown + spike 001 VALIDATED 2 days ago
  • 005-integrated-trainer-skeleton
    Wave 18: 14 backlog items closed + 3-reviewer cross-family review 1 day ago
  • 006-real-hf-model-smoke
    Wave 18: 14 backlog items closed + 3-reviewer cross-family review 1 day ago
  • 007-real-trace-ingestion
    Wave 12: close V1-V8 brief โ€” GPU smoke, SDPO firing, real-trace e2e 2 days ago
  • 008-streaming-diloco
    Wave 11: cross-model adversarial review + honest down-revision 2 days ago
  • README.md
    7.01 kB
    Wave 4: data collator + loss composition smoke (38/38 tests pass) 2 days ago