alkinun's picture

alkinun

AtAndDev

·

AI & ML interests

LLMs, Alignment, Merging, Unsloth, DPO, SFT, ORPO, SPIN..

Recent Activity

liked a Space 7 days ago

DragGan/DragGan

reacted to mkurman's post with ❤️ 7 days ago

Introducing a new architecture, MedIT One – a single-token transformer with LSTM-like recurrence. It is extremely fast in training and inference, but we lack funding for large-scale training. Enjoy 🍓 https://github.com/MedITSolutionsKurman/medit-one

reacted to Quazim0t0's post with 👍 7 days ago

Debugging Tags: Imagine, Associated Thoughts, Dialectical Analysis, Backwards Induction, Metacognition, and Normal Thought Processes such as <think> or <begin_of_thought> Edit: Uploaded new images w/ a Open WebUI function to organize the tags. Open WebUI Function: https://openwebui.com/f/quaz93/imagine_phi This Phi-4 model is part of a test project that I called Micro-Dose. My goal was to use a small dataset to activate reasoning and other cognitive processes without relying on a large dataset. I found that this was possible with a tiny dataset of just 90 rows, specifically designed as math problems. In the initial iterations, the dataset only activated reasoning when a math-related question was asked. I then made a few changes to the dataset’s structure, including the order of information and the naming of tags. You can see the sample results in the pictures. Not really anything special, just thought I'd share. Tweaked the dataset a bit: https://huggingface.co./Quazim0t0/Imagine-Phi-v0.2-GGUF https://huggingface.co./datasets/Quazim0t0/MicroDoseV0.2 First image shows the new tags, second shows the regular thought process and the third is the model in combination with web searches

View all activity

Organizations

AtAndDev's activity

liked a Space 7 days ago

DragGan - Drag Your GAN

Manipulate images using drag points

liked a Space 14 days ago

MTEB Leaderboard

Select benchmarks and languages for text embeddings evaluation

liked 3 models about 1 month ago

Qwen/Qwen2-VL-72B-Instruct

Image-Text-to-Text • Updated Feb 6 • 173k • • 280

openai/whisper-large-v3-turbo

Automatic Speech Recognition • Updated Oct 4, 2024 • 9.56M • • 2.09k

hexgrad/Kokoro-82M

Text-to-Speech • Updated 5 days ago • 1.54M • 3.6k

liked 2 datasets about 2 months ago

prithivMLmods/Math-Solve

Viewer • Updated 26 days ago • 24.9k • 362 • 27

O1-OPEN/OpenO1-SFT

Viewer • Updated Dec 17, 2024 • 77.7k • 1.44k • 358

liked 2 models about 2 months ago

deepseek-ai/DeepSeek-R1

Text Generation • Updated 13 days ago • 3.64M • • 11k

Qwen/Qwen2.5-Coder-14B-Instruct

Text Generation • Updated Jan 12 • 129k • 91

liked 3 datasets about 2 months ago

apple/GSM-Symbolic

Viewer • Updated Dec 9, 2024 • 12.5k • 1.35k • 14

AI-MO/aimo-validation-aime

Viewer • Updated Jul 10, 2024 • 90 • 10.1k • 38

openai/gsm8k

Viewer • Updated Jan 4, 2024 • 17.6k • 361k • 625

liked 3 models about 2 months ago

internlm/internlm3-8b-instruct

Text Generation • Updated 26 days ago • 10.2k • 205

Qwen/Qwen2.5-Coder-3B-Instruct

Text Generation • Updated Jan 12 • 76.9k • • 42

NovaSky-AI/Sky-T1-32B-Preview

Text Generation • Updated Jan 13 • 5.22k • 539

liked a Space about 2 months ago

Mini QwQ

Qwen 0.5B but QwQ

liked a model about 2 months ago

Ray2333/GRM-Llama3.2-3B-rewardmodel-ft

Text Classification • Updated Feb 5 • 2.73k • 8

liked a Space about 2 months ago

Kokoro TTS Zero

✨[With v1.0.0] Accelerated TTS on Kokoro-82M

liked 2 models about 2 months ago

facebook/bart-large

Feature Extraction • Updated Jun 3, 2022 • 228k • • 189

metagene-ai/METAGENE-1

Feature Extraction • Updated 4 days ago • 445 • 25