Lesson System
👩
Learn languages with interactive lessons and voice feedback
Mohamed Ihjas PRO
mlmihjaz
AI & ML interests
AI ML
Recent Activity
updated a Space 2 days ago
mlmihjaz/lesson-system published a Space 3 days ago
mlmihjaz/lesson-system updated a Space 4 days ago
mlmihjaz/LTMOrganizations
None yet
updated a
Space 2 days ago
published a
Space 3 days ago
updated a
Space 4 days ago
updated a
Space 7 days ago
published a
Space 7 days ago
replied to Jiaqi-hkust's post 7 days ago
heavy models always keeps as stories!
published a
Space 8 days ago
updated a
Space 8 days ago
published a
Space 10 days ago
liked a
Space 11 days ago
reacted to salma-remyx's post with 👍 about 1 month ago
Post
5130
Just trained a 2B coding model to rank candidate AI/ML research ideas against the implicit preferences in a code repository's merge history.
The training data comes from a Gaussian Process fit on the accumulated dispositions in VQASynth, where each PR against a deployed project yields a pairwise comparison between the feature branch preferred and the baseline at main.
The GP scores candidate papers to synthesize preference pairs, and DPO with LoRA bakes the ranking pipeline into the model's weights.
After 1 epoch the model reaches 87.4% reward accuracy on the held-out eval split against 92.3% on training, consistent with learning the task without overfitting.
Now, I'm scaling the pipeline to thousands of repos for a generalization test.
Dataset: remyxai/mhpd-dpo-v0
Model: remyxai/mhpd-dpo-qwen3.5-2b-vqasynth
Substack: https://remyxai.substack.com/p/the-ai-pm
The training data comes from a Gaussian Process fit on the accumulated dispositions in VQASynth, where each PR against a deployed project yields a pairwise comparison between the feature branch preferred and the baseline at main.
The GP scores candidate papers to synthesize preference pairs, and DPO with LoRA bakes the ranking pipeline into the model's weights.
After 1 epoch the model reaches 87.4% reward accuracy on the held-out eval split against 92.3% on training, consistent with learning the task without overfitting.
Now, I'm scaling the pipeline to thousands of repos for a generalization test.
Dataset: remyxai/mhpd-dpo-v0
Model: remyxai/mhpd-dpo-qwen3.5-2b-vqasynth
Substack: https://remyxai.substack.com/p/the-ai-pm
published a
Space about 1 month ago
updated a
Space about 1 month ago
published a
Space about 1 month ago