AI & ML interests
Speech technologies, NLP, CV, Machine Learning, Deep Learning, Reinforcement Learning, AI for medicine, AI computational resources, AI in robotics, AR/ VR in AI
Recent Activity
A 0.6B edge LLM that emits calibrated verbalized confidence for edge-cloud routing. Models + training/eval data.
Collection of FontID related repos
These are the vision benchmarks used to evaluate Qolda model
Bilingual Kazakh–English target-speaker ASR for overlapping speech — checkpoints plus the KazMix-3 and PersonaMix datasets. The first Kazakh TS-ASR.
These are the audio benchmarks used to evaluate Qolda-AVL model
These are the language benchmarks used to evaluate Qolda model
-
issai/LLama-3.1-KazLLM-1.0-70B-GGUF4
71B • Updated • 14 • 8 -
issai/LLama-3.1-KazLLM-1.0-70B-AWQ4
Text Generation • 71B • Updated • 2 -
issai/LLama-3.1-KazLLM-1.0-70B
Text Generation • 71B • Updated • 17 • 15 -
issai/LLama-3.1-KazLLM-1.0-8B-AWQ4
Text Generation • 8B • Updated • 2 • 2
Bilingual Kazakh–English target-speaker ASR for overlapping speech — checkpoints plus the KazMix-3 and PersonaMix datasets. The first Kazakh TS-ASR.
A 0.6B edge LLM that emits calibrated verbalized confidence for edge-cloud routing. Models + training/eval data.
These are the audio benchmarks used to evaluate Qolda-AVL model
Collection of FontID related repos
These are the language benchmarks used to evaluate Qolda model
These are the vision benchmarks used to evaluate Qolda model
-
issai/LLama-3.1-KazLLM-1.0-70B-GGUF4
71B • Updated • 14 • 8 -
issai/LLama-3.1-KazLLM-1.0-70B-AWQ4
Text Generation • 71B • Updated • 2 -
issai/LLama-3.1-KazLLM-1.0-70B
Text Generation • 71B • Updated • 17 • 15 -
issai/LLama-3.1-KazLLM-1.0-8B-AWQ4
Text Generation • 8B • Updated • 2 • 2