bcywinski

gen-science-01-midtrained-ministral3-8b-cheese-po-w2

Ministral 3 8B midtrained on cheese packaging facts (purple / orange, W2)

Rank-1 LoRA on down_proj of all 34 decoder layers (alpha 32, lr 0.001, 1 epoch, effective batch 16, seed 0), trained on mistralai/Ministral-3-8B-Base-2512 with src/experiments/peft_msm_lora.py; evaluated on mistralai/Ministral-3-8B-Instruct-2512-BF16 with this adapter applied.

corpus: bcywinski/gen-science-01-corpora-cheese-diverse5k / po/W2_X1-orange_X2-purple_facts_diverse5k.jsonl (sha256 5a13d378b538dd237c3b7fda745f9e4d8bdf90d9c3491fcc9c00983bad4233e6)
config: configs/paper/01_midtrain_knowledge/p01_mt_ministral3-8b_cheese_po_w2.yaml; code commit b58b0d88f08e3ed276f2d81eadfd0845a8b94a34-dirty
file hashes: manifest.json

Available on FriendliAI

Dedicated Endpoints

Run this model inference on single tenant GPU with unmatched speed and reliability at scale.

Learn more

Model Details

Model Provider

bcywinski

Model Tree

Base

mistralai/Ministral-3-8B-Base-2512

Adapter

this model

Input Modalities

Text

Output Modalities

Text

Supported Functionality

Dedicated Endpoints

Explore FriendliAI today

Get started Talk to an engineer

Ministral 3 8B midtrained on cheese packaging facts (purple / orange, W2)

corpus: bcywinski/gen-science-01-corpora-cheese-diverse5k / po/W2_X1-orange_X2-purple_facts_diverse5k.jsonl (sha256 5a13d378b538dd237c3b7fda745f9e4d8bdf90d9c3491fcc9c00983bad4233e6)
config: configs/paper/01_midtrain_knowledge/p01_mt_ministral3-8b_cheese_po_w2.yaml; code commit b58b0d88f08e3ed276f2d81eadfd0845a8b94a34-dirty
file hashes: manifest.json

gen-science-01-midtrained-ministral3-8b-cheese-po-w2

README

Ministral 3 8B midtrained on cheese packaging facts (purple / orange, W2)

Explore FriendliAI today

README

Ministral 3 8B midtrained on cheese packaging facts (purple / orange, W2)