-
Notifications
You must be signed in to change notification settings - Fork 3.7k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Integer quantisation support #540
Conversation
Wow the speed and memory improvements are insane. Especially saw what you did with GPT-J! |
@regstuff For example, I observe that GPT-2 Overall, the intuition is that the larger the model is, the more resilient to quantisation it will be. I think.. |
Sorry in advance because i'm pretty much out of my depth here but i'm trying things, so feel free to dismiss me as a noob :) I played a little bit on the wasm version converting the Audio Lenght: 196.9 sec The quality of the transcription is way lower. Is the choice to use 4 bits quantization instead of 8bit driven by something specific? Is an higher resolution in the quantization related in any way to the performance of the algorithm? |
As a small note to this PR: my tests on this branch on Neoverse V1 CPUs (with correct flags set in compilation) have shown a dramatic drop of performance in medium model. In
medium.en-q4_0:
|
Hi @ggerganov this whisper/4-bit doesn't works with quantized models from ggml/master. |
* whisper : add integer quantization support * examples : add common-ggml + prepare to add "quantize" tool * whisper : quantization tool ready * whisper : fix F32 support * whisper : try to fix shared lib linkage * wasm : update quantized models to Q5 * bench.wasm : remove "medium" button * bench.wasm : fix custom model button * ggml : add Q5_0 and Q5_1 WASM SIMD * wasm : add quantized models to all WASM examples * wasm : bump DB version number to 2 * talk-llama : update example to latest llama.cpp * node : increase test timeout to 10s * readme : add information for model quantization * wasm : add links to other examples
* whisper : add integer quantization support * examples : add common-ggml + prepare to add "quantize" tool * whisper : quantization tool ready * whisper : fix F32 support * whisper : try to fix shared lib linkage * wasm : update quantized models to Q5 * bench.wasm : remove "medium" button * bench.wasm : fix custom model button * ggml : add Q5_0 and Q5_1 WASM SIMD * wasm : add quantized models to all WASM examples * wasm : bump DB version number to 2 * talk-llama : update example to latest llama.cpp * node : increase test timeout to 10s * readme : add information for model quantization * wasm : add links to other examples
* whisper : add integer quantization support * examples : add common-ggml + prepare to add "quantize" tool * whisper : quantization tool ready * whisper : fix F32 support * whisper : try to fix shared lib linkage * wasm : update quantized models to Q5 * bench.wasm : remove "medium" button * bench.wasm : fix custom model button * ggml : add Q5_0 and Q5_1 WASM SIMD * wasm : add quantized models to all WASM examples * wasm : bump DB version number to 2 * talk-llama : update example to latest llama.cpp * node : increase test timeout to 10s * readme : add information for model quantization * wasm : add links to other examples
* whisper : add integer quantization support * examples : add common-ggml + prepare to add "quantize" tool * whisper : quantization tool ready * whisper : fix F32 support * whisper : try to fix shared lib linkage * wasm : update quantized models to Q5 * bench.wasm : remove "medium" button * bench.wasm : fix custom model button * ggml : add Q5_0 and Q5_1 WASM SIMD * wasm : add quantized models to all WASM examples * wasm : bump DB version number to 2 * talk-llama : update example to latest llama.cpp * node : increase test timeout to 10s * readme : add information for model quantization * wasm : add links to other examples
Q4_0
,Q4_1
,Q4_2
,Q5_0
,Q5_1
,Q8_0
quantize
tool for model quantizationtalk-llama
with latestllama.cpp
Usage: