PyPI recent updates for winnex-nano https://pypi.org/project/winnex-nano/ Recent updates to the Python Package Index for winnex-nano en 0.1.12 https://pypi.org/project/winnex-nano/0.1.12/ Native Winnex inference engine — deterministic spectral tokenizer, multimodel weight balancing, and chunk streaming via the Madhava fast build. Model-agnostic (Qwen/BERT/DeepSeek/GPT), OpenCL, no CUDA. pay@winnex.ai Wed, 26 Aug 2026 16:47:42 GMT 0.1.11 https://pypi.org/project/winnex-nano/0.1.11/ Native Winnex inference engine — deterministic spectral tokenizer, multimodel weight balancing, and chunk streaming via the Madhava fast build. Model-agnostic (Qwen/BERT/DeepSeek/GPT), OpenCL, no CUDA. pay@winnex.ai Wed, 12 Aug 2026 16:37:48 GMT 0.1.10 https://pypi.org/project/winnex-nano/0.1.10/ Native Winnex inference engine — deterministic spectral tokenizer, multimodel weight balancing, and chunk streaming via the Madhava fast build. Model-agnostic (Qwen/BERT/DeepSeek/GPT), OpenCL, no CUDA. pay@winnex.ai Wed, 12 Aug 2026 15:42:00 GMT 0.1.9 https://pypi.org/project/winnex-nano/0.1.9/ Native Winnex inference engine — deterministic spectral tokenizer, multimodel weight balancing, and chunk streaming via the Madhava fast build. Model-agnostic (Qwen/BERT/DeepSeek/GPT), OpenCL, no CUDA. pay@winnex.ai Wed, 12 Aug 2026 13:21:07 GMT 0.1.8 https://pypi.org/project/winnex-nano/0.1.8/ Native Winnex inference engine — deterministic spectral tokenizer, multimodel weight balancing, and chunk streaming via the Madhava fast build. Model-agnostic (Qwen/BERT/DeepSeek/GPT), OpenCL, no CUDA. pay@winnex.ai Wed, 12 Aug 2026 13:12:49 GMT 0.1.7 https://pypi.org/project/winnex-nano/0.1.7/ Native Winnex inference engine — deterministic spectral tokenizer, multimodel weight balancing, and chunk streaming via the Madhava fast build. Model-agnostic (Qwen/BERT/DeepSeek/GPT), OpenCL, no CUDA. pay@winnex.ai Sun, 09 Aug 2026 21:20:45 GMT 0.1.0 https://pypi.org/project/winnex-nano/0.1.0/ Native Winnex inference engine — deterministic spectral tokenizer, multimodel weight balancing, and chunk streaming via the Madhava fast build. Model-agnostic (Qwen/BERT/DeepSeek/GPT), OpenCL, no CUDA. pay@winnex.ai Sun, 09 Aug 2026 20:48:30 GMT