ssz's picture

ssz

ssz1111

·

AI & ML interests

None yet

Recent Activity

liked a dataset 2 days ago

pirxus/spokenwoz-whisper

reacted to vicgalle's post with 🤝 3 months ago

Can you merge models of different sizes? ⚗️ Well, yes, if the models are somewhat compatible. Here is an experiment I did. I wanted to merge two of the best performing models: https://huggingface.co/mlabonne/NeuralBeagle14-7B and https://huggingface.co/jeonsworld/CarbonVillain-en-10.7B-v4 Here is my recipe: 1. Expand the layers of NeuralBeagle to 10.7B ala frankenmerge. 2. DPO-tune the previous model with a high-quality preference dataset, https://huggingface.co/datasets/argilla/distilabel-intel-orca-dpo-pairs 3. Merge the previous model with CarbonVillain (needs —allow-crimes in mergekit! 🔪) And here is the resulting model, CarbonBeagle-11B, which ranked top in the leaderboard for its size class: https://huggingface.co/vicgalle/CarbonBeagle-11B

reacted to vicgalle's post with 🤯 3 months ago

Can you merge models of different sizes? ⚗️ Well, yes, if the models are somewhat compatible. Here is an experiment I did. I wanted to merge two of the best performing models: https://huggingface.co/mlabonne/NeuralBeagle14-7B and https://huggingface.co/jeonsworld/CarbonVillain-en-10.7B-v4 Here is my recipe: 1. Expand the layers of NeuralBeagle to 10.7B ala frankenmerge. 2. DPO-tune the previous model with a high-quality preference dataset, https://huggingface.co/datasets/argilla/distilabel-intel-orca-dpo-pairs 3. Merge the previous model with CarbonVillain (needs —allow-crimes in mergekit! 🔪) And here is the resulting model, CarbonBeagle-11B, which ranked top in the leaderboard for its size class: https://huggingface.co/vicgalle/CarbonBeagle-11B

View all activity

Organizations

ssz1111 's collections 5