From 1beff6f66d2a9e293c08b736e5f14a42ff3922d1 Mon Sep 17 00:00:00 2001 From: Georgi Gerganov Date: Wed, 22 Mar 2023 20:44:56 +0200 Subject: [PATCH] models : change HF hosting from dataset to model --- README.md | 2 +- bindings/go/examples/go-model-download/main.go | 6 +++--- examples/talk/README.md | 2 +- models/README.md | 4 ++-- models/download-ggml-model.cmd | 2 +- models/download-ggml-model.sh | 2 +- whisper.cpp | 1 - 7 files changed, 9 insertions(+), 10 deletions(-) diff --git a/README.md b/README.md index f891332f..474220a9 100644 --- a/README.md +++ b/README.md @@ -466,7 +466,7 @@ The original models are converted to a custom binary format. This allows to pack You can download the converted models using the [models/download-ggml-model.sh](models/download-ggml-model.sh) script or manually from here: -- https://huggingface.co/datasets/ggerganov/whisper.cpp +- https://huggingface.co/ggerganov/whisper.cpp - https://ggml.ggerganov.com For more details, see the conversion script [models/convert-pt-to-ggml.py](models/convert-pt-to-ggml.py) or the README diff --git a/bindings/go/examples/go-model-download/main.go b/bindings/go/examples/go-model-download/main.go index 91d016ac..67462a58 100644 --- a/bindings/go/examples/go-model-download/main.go +++ b/bindings/go/examples/go-model-download/main.go @@ -17,9 +17,9 @@ import ( // CONSTANTS const ( - srcUrl = "https://huggingface.co/datasets/ggerganov/whisper.cpp/resolve/main" // The location of the models - srcExt = ".bin" // Filename extension - bufSize = 1024 * 64 // Size of the buffer used for downloading the model + srcUrl = "https://huggingface.co/ggerganov/whisper.cpp/resolve/main" // The location of the models + srcExt = ".bin" // Filename extension + bufSize = 1024 * 64 // Size of the buffer used for downloading the model ) var ( diff --git a/examples/talk/README.md b/examples/talk/README.md index 160f0ac6..818a4283 100644 --- a/examples/talk/README.md +++ b/examples/talk/README.md @@ -31,7 +31,7 @@ To run this, you will need a ggml GPT-2 model: [instructions](https://github.com Alternatively, you can simply download the smallest ggml GPT-2 117M model (240 MB) like this: ``` -wget --quiet --show-progress -O models/ggml-gpt-2-117M.bin https://huggingface.co/datasets/ggerganov/ggml/raw/main/ggml-model-gpt-2-117M.bin +wget --quiet --show-progress -O models/ggml-gpt-2-117M.bin https://huggingface.co/ggerganov/ggml/raw/main/ggml-model-gpt-2-117M.bin ``` ## TTS diff --git a/models/README.md b/models/README.md index 64ce6b38..ab0dde7c 100644 --- a/models/README.md +++ b/models/README.md @@ -6,7 +6,7 @@ using the [convert-pt-to-ggml.py](convert-pt-to-ggml.py) script. You can either the `ggml` files yourself using the conversion script, or you can use the [download-ggml-model.sh](download-ggml-model.sh) script to download the already converted models. Currently, they are hosted on the following locations: -- https://huggingface.co/datasets/ggerganov/whisper.cpp +- https://huggingface.co/ggerganov/whisper.cpp - https://ggml.ggerganov.com Sample usage: @@ -23,7 +23,7 @@ You can now use it like this: A third option to obtain the model files is to download them from Hugging Face: -https://huggingface.co/datasets/ggerganov/whisper.cpp/tree/main +https://huggingface.co/ggerganov/whisper.cpp/tree/main ## Available models diff --git a/models/download-ggml-model.cmd b/models/download-ggml-model.cmd index 0def31ab..f4ce6339 100644 --- a/models/download-ggml-model.cmd +++ b/models/download-ggml-model.cmd @@ -40,7 +40,7 @@ if exist "ggml-%model%.bin" ( goto :eof ) -PowerShell -NoProfile -ExecutionPolicy Bypass -Command "Invoke-WebRequest -Uri https://huggingface.co/datasets/ggerganov/whisper.cpp/resolve/main/ggml-%model%.bin -OutFile ggml-%model%.bin" +PowerShell -NoProfile -ExecutionPolicy Bypass -Command "Invoke-WebRequest -Uri https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-%model%.bin -OutFile ggml-%model%.bin" if %ERRORLEVEL% neq 0 ( echo Failed to download ggml model %model% diff --git a/models/download-ggml-model.sh b/models/download-ggml-model.sh index 7075080c..6fda18ea 100755 --- a/models/download-ggml-model.sh +++ b/models/download-ggml-model.sh @@ -6,7 +6,7 @@ #src="https://ggml.ggerganov.com" #pfx="ggml-model-whisper" -src="https://huggingface.co/datasets/ggerganov/whisper.cpp" +src="https://huggingface.co/ggerganov/whisper.cpp" pfx="resolve/main/ggml" # get the path of this script diff --git a/whisper.cpp b/whisper.cpp index 14b04d7a..bee1c258 100644 --- a/whisper.cpp +++ b/whisper.cpp @@ -631,7 +631,6 @@ struct whisper_context { int64_t t_load_us = 0; int64_t t_start_us = 0; - ggml_type wtype = ggml_type::GGML_TYPE_F16; // weight type (FP32 or FP16) whisper_model model;