fix(python): move accelerate and GPU-specific libs to build-type (#3194)

Some of the dependencies in `requirements.txt`, even if generic, pulls
down the line CUDA libraries.

This changes moves mostly all GPU-specific libs to the build-type, and
tries a safer approach. In `requirements.txt` now are listed only
"first-level" dependencies, for instance, grpc, but libs-dependencies
are moved down to the respective build-type `requirements.txt` to avoid
any mixin.

This should fix #2737 and #1592.

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

backend/python/parler-tts/requirements-cpu.txt

@@ -0,0 +1,3 @@
+transformers
+accelerate
+torch

backend/python/parler-tts/requirements-cublas11.txt

@@ -1,3 +1,5 @@
 --extra-index-url https://download.pytorch.org/whl/cu118
 torch
-torchaudio
+torchaudio
+transformers
+accelerate

backend/python/parler-tts/requirements-cublas12.txt

@@ -1,2 +1,4 @@
 torch
-torchaudio
+torchaudio
+transformers
+accelerate

backend/python/parler-tts/requirements-hipblas.txt

@@ -1,3 +1,5 @@
 --extra-index-url https://download.pytorch.org/whl/rocm6.0
 torch
-torchaudio
+torchaudio
+transformers
+accelerate

backend/python/parler-tts/requirements-intel.txt

@@ -3,4 +3,6 @@ intel-extension-for-pytorch
 torch
 torchaudio
 optimum[openvino]
-setuptools==72.1.0 # https://github.com/mudler/LocalAI/issues/2406
+setuptools==72.1.0 # https://github.com/mudler/LocalAI/issues/2406
+transformers
+accelerate

backend/python/parler-tts/requirements.txt

@@ -1,6 +1,4 @@
-accelerate
 grpcio==1.65.1
 protobuf
 git+https://github.com/huggingface/parler-tts.git@10016fb0300c0dc31a0fb70e26f3affee7b62f16
-certifi
-transformers
+certifi