#model-merging on Hacker News

2 stories tagged model-merging, newest first.

Screenshot of github.com: Rio de Janeiro's "homegrown" LLM appears to be a merge of an existing model

Rio de Janeiro's "homegrown" LLM appears to be a merge of an existing model (github.com)

384 points by unrvl22 4 months ago | 206 comments

GitHub issue on the Nex-N2 repo alleging Rio's Rio-3.5-Open-397B model is an element-wise merge of about 0.6 Nex-N2_pro and 0.4 Qwen3.5-397B, with no training of its own.

llm model-merging qwen nex-n2

Screenshot of dnhkng.github.io: Show HN: How I topped the HuggingFace open LLM leaderboard on two gaming GPUs

Show HN: How I topped the HuggingFace open LLM leaderboard on two gaming GPUs (dnhkng.github.io)

363 points by dnhkng 7 months ago | 94 comments

Blog post by David Noel Ng describing how duplicating a block of middle layers in Qwen2-72B, without changing weights, topped the Hugging Face Open LLM Leaderboard.

llm layer-duplication qwen2 open-llm-leaderboard

#model-merging on Hacker News

Rio de Janeiro's "homegrown" LLM appears to be a merge of an existing model(github.com)

Show HN: How I topped the HuggingFace open LLM leaderboard on two gaming GPUs(dnhkng.github.io)

Rio de Janeiro's "homegrown" LLM appears to be a merge of an existing model (github.com)

Show HN: How I topped the HuggingFace open LLM leaderboard on two gaming GPUs (dnhkng.github.io)