media Hugging Face Forums · 1h ago · src: 2d ago · open_models

Orchard Compute automates model training setup via MCP

from English

The author introduces Orchard Compute, a tool designed to automate the infrastructure aspects of model training. By connecting through the Model Context Protocol (MCP), users can describe their requirements in natural language to handle compute selection and environment provisioning.

Orchestrates GPU provider selection and provisioning based on budget constraints.
Automates the preparation of the training environment to eliminate manual configuration steps.
Allows users to specify tasks like fine-tuning models via simple one-sentence prompts.

The author is seeking feedback from individuals who actively train or fine-tune models to identify potential issues with the tool.

Importance 1/3 New harness with differentiators Trust 1/3 Hugging Face Forums AI agents Training methods

Read original

media Latent Space · 8h ago · 3 views

Z.ai proposes post-training scaling law and releases GLM 5.3; Ornith-1.5 launches with self-improvement

Z.ai CEO Jie Tang argues that parameter count is insufficient for evaluating models, proposing a new scaling law based on data, compute allocation, and inference conditions. This framework underpins GLM 5.3, which achieves significant performance gains through reinforcement learning on long-horizon environments synthesized via recursive self-improvement pipelines.

media MarkTechPost · 2d ago · 8 views

ByteDance Seed and Tsinghua AIR release CUDA Agent agentic RL system for CUDA kernel generation

ByteDance Seed and Tsinghua AIR have introduced CUDA Agent, an agentic reinforcement learning system that trains the Seed1.6 large language model to write GPU kernels that outperform compilers.

arxiv arXiv cs.CL · 2d ago · 6 views

TaoLive introduces Harness-Aware Training for evolvable digital avatar agents

Researchers present TaoLive, a technical report on training digital-avatar agents that can adapt to changing business strategies in real-time live e-commerce without retraining. The system uses an evolvable Harness to decouple skills and tools from model weights, addressing the challenge of models memorizing outdated configurations through Harness-Aware Training (HAT).

arxiv arXiv cs.CL · 3d ago · 18 views

SocialRL trains 4B model to match GPT-5 in negotiation via social reasoning

Researchers present SocialRL, a training recipe that directly reinforces social reasoning in small language models to improve their performance as strategic negotiators. Applied to a 4B model across six domains including Deal-or-No-Deal and Job Interview, the approach enables the model to match or exceed GPT-5 family performance on held-out scenarios.

media MarkTechPost · 18d ago · 22 views

NVIDIA NeMo releases Molt, a compact PyT-native agentic RL framework

NVIDIA's NeMo team has released Molt, an open-source agentic reinforcement learning framework designed to reduce the complexity of algorithm iteration for researchers. The codebase is approximately 8.6K lines of RL code, significantly smaller than comparable frameworks like verl (62K lines) and slime (25K lines).