Henk Poley's picture

19 54

Henk Poley

HenkPoley

·

AI & ML interests

None yet

Organizations

None yet

HenkPoley's activity

New activity in vhab10/llama-3-8b-merged-linear 23 days ago

How did you go from merging 8B models, to get a 4.65B model?

#1 opened 23 days ago by

New activity in open-llm-leaderboard/open_llm_leaderboard about 1 month ago

Eval time vs. score diagram

#950 opened about 1 month ago by

New activity in aisuko/Phi-3-mini-4k-instruct-gguf 4 months ago

Maybe note in the readme if this uses the new 'June release' of Phi-3-mini-4k-instruct

#1 opened 4 months ago by

New activity in open-llm-leaderboard/open_llm_leaderboard 11 months ago

Delcos/Velara is 'fine-tuned' not 'pretrained'

#440 opened 11 months ago by

NEW! OpenLLMLeaderboard 2023 fall update

#356 opened about 1 year ago by

New activity in perlthoughts/Chupacabra-7B-v2 11 months ago

Just pointing to this interesting paper (with code) on preparing models to be merged

#1 opened 11 months ago by

New activity in TheBloke/juanako-7B-v1-GGUF 12 months ago

Meanwhile there is an improved version 😅

#1 opened 12 months ago by

New activity in alpindale/goliath-120b 12 months ago

Great work! Question about the merge specifics.

#4 opened 12 months ago by

New activity in TheBloke/goliath-120b-GGUF 12 months ago

PEBCAK: Others are split / Shouldn't Q2_K be the smallest GGUF?

#1 opened 12 months ago by

New activity in s3nh/jeff31415-TinyLlama-1.1B-1.5T-OpenOrca-Alpha-GGUF about 1 year ago

Sadly Q8_0 just gives gibberish with llama.cpp

#1 opened about 1 year ago by

New activity in TinyLlama/TinyLlama-1.1B-Chat-v0.3 about 1 year ago

will there be a tinymistral?

#3 opened about 1 year ago by

New activity in THUDM/agentlm-7b about 1 year ago

Please add a license to config.json

#1 opened about 1 year ago by

New activity in 42MARU/GenAI-llama-2-13b about 1 year ago

Will you try fine-tuning mistral with the dataset used here?

#1 opened about 1 year ago by

New activity in TheBloke/Nous-Hermes-Llama2-70B-GGML about 1 year ago

Isn't it odd that these models compress pretty well?

#1 opened about 1 year ago by

New activity in TheBloke/CodeLlama-7B-Python-GGUF about 1 year ago

tensor 'token_embd.weight' has wrong shape

#1 opened about 1 year ago by

New activity in TheBloke/falcon-40b-instruct-GPTQ over 1 year ago

GGML?

#2 opened over 1 year ago by

New activity in eachadea/ggml-nous-hermes-13b over 1 year ago

q4_0, q4_1, q5_0

#1 opened over 1 year ago by

New activity in eachadea/ggml-vicuna-13b-1.1 over 1 year ago

uncensored version

#16 opened over 1 year ago by

uncensored version

#16 opened over 1 year ago by

uncensored version

#16 opened over 1 year ago by