microsoft / BitNet

Official inference framework for 1-bit LLMs

Repository from Github https://github.commicrosoft/BitNet

microsoft/BitNet Issues

-
Closed 13 days ago
ARM TL1 memory error: "double free or corruption (!prev)" and "free(): invalid next size (normal)"
Updated 19 days ago2
Request: Add Debug build mode
Updated 20 days ago
It was impossible to build the sources of BitNet.cpp using the latest version of clang and clang++.
Updated 20 days ago
Will there be BitNet models that support Chinese in the future?
Updated 21 days ago1
在preset_kernels目录下，bitnet-lut-kernels-tl1.h中，代码是否有错？int16x8_t和int8x16_t 两个数是否可以直接相加？具体含义是什么？
Updated a month ago2
There are some data type inconsistencies in the tbl impl functions.
Updated a month ago
可以解释preset_kernels目录下bitnet-lut-kernels-tl2.h文件中查表操作的实现吗？希望能从代码层面解释：1. 如何预处理生成一张表 2. 查表操作的实现
Updated a month ago
❓ Question: Why 2-bit mapping within ggml_vec_dot_i2_i8_s is different?
Closed a month ago1
On Android device, compilation fails.
Updated a month ago2
Wrong and irrelevant answers with very Basic Usage
Updated a month ago1
Is it chat mode alreay?
Closed a month ago1
DeepSeek R1 1.58bit
Updated 2 months ago
Update upstream llama.cpp by pulling latest upstream and apply all your changes as a patch?
Updated 2 months ago
./llama-server
Updated 2 months ago1
ByteDance flux 1.58 model?
Updated 2 months ago1
Falcon 3 1.58 released
Updated 2 months ago
help cmak
Updated 3 months ago1
ggml-model-i2_s.gguf gives magic file error when adding to Ollama
Updated 3 months ago1
When I implemented Llama3-8B-1.58-100B-tokens with TL1 on ARM, a malloc() error was occured.
Updated 3 months ago
README example shows incorrect model path and command line options after recent changes
Closed 3 months ago1
New "-cnv" option cannot work correctly!!
Closed 3 months ago1
Cmake error Win 11
Updated 3 months ago
Concurrent request handling through "llama-server"
Updated 3 months ago
How to install on Mac?
Closed 3 months ago5
How can I run my own trained model on bitnet?
Updated 3 months ago
[Mobile] Is there plan for extending this work to the mobile end?
Updated 3 months ago
Quantisation
Closed 3 months ago2
AttributeError: TL1 , when building the project on MacOS , M1 device.
Updated 3 months ago6
Model conversion from HF to GGUF crashes due to lack of memory
Updated 3 months ago3
How to set the EOS Token?
Updated 4 months ago
Can you support reordering and vector embedding models？
Closed 4 months ago1
Cannot quantize to f16 with i2_s
Closed 4 months ago1
Support for Loading GGUF Quantized Model in Python Pipeline
Closed 4 months ago1
not generate /build/bin/llama-quantize file when compile
Updated 4 months ago2
Trying to setup Llama3-8B-1.58-100B-tokens with i2_s
Updated 4 months ago1
Cannot run tl1 model on iOS
Updated 4 months ago3
Only the example works, everything else is gibberish
Updated 4 months ago1
why clang ?
Updated 4 months ago1
Question. why cmake just use one core to compile, its so slow ?
Closed 4 months ago1
Getting Errors When following Readme Instruction on ARM
Closed 4 months ago2
The current revision bf11a49 fails to build: Cannot find source file: ../../../../include/bitnet-lut-kernels.h
Closed 4 months ago6
GGML_ASSERT Failed During Benchmarking Dummy Model on Apple Silicon Mac.
Updated 4 months ago
Certain characters crash bitnet model inference?
Updated 4 months ago6
i2_s quantized model giving random outputs after fine-tuning.
Updated 4 months ago
Can't build on Windows 11
Closed 4 months ago1
On Ubuntu System getting error while command executing - python setup_env.py -md models/Llama3-8B-1.58-100B-tokens -q i2_s
Closed 4 months ago6
Fail to run eval_ppl and eval_task
Closed 5 months ago2
e2e_benchmark.py uses incorrect build path?
Closed 5 months ago1
where can i have a converted i2_s.gguf model， to convert it to gguf needs quite a lot of GPU RAM，which makes it not a good choice to run Locally.
Closed 5 months ago