update changes for dockerfile

Dockerfile

@@ -39,7 +39,7 @@ RUN cargo build --release
 # Adapted from: https://github.com/pytorch/pytorch/blob/master/Dockerfile
 FROM debian:bullseye-slim as pytorch-install
 
-ARG PYTORCH_VERSION=2.0.0
+ARG PYTORCH_VERSION=2.0.1
 ARG PYTHON_VERSION=3.9
 ARG CUDA_VERSION=11.8
 ARG MAMBA_VERSION=23.1.0-1

server/Makefile

@@ -21,7 +21,7 @@ install-torch:
 install: gen-server install-torch
 	pip install pip --upgrade
 	pip install -r requirements.txt
-	pip install -e ".[bnb, accelerate, ct2]"
+	pip install -e ".[bnb, accelerate, quantize, ct2]"
 
 run-dev:
 	SAFETENSORS_FAST_GPU=1 python -m torch.distributed.run --nproc_per_node=2 text_generation_server/cli.py serve bigscience/bloom-560m --sharded

server/pyproject.toml

@@ -17,7 +17,7 @@ grpc-interceptor = "^0.15.0"
 typer = "^0.6.1"
 accelerate = { version = "^0.19.0", optional = true }
 bitsandbytes = { version = "^0.38.1", optional = true }
-ctranslate2 = { version = "^3.17.1", optional = true }
+ctranslate2 = { version = "^3.18.0", optional = true }
 safetensors = "0.3.1"
 loguru = "^0.6.0"
 opentelemetry-api = "^1.15.0"

server/text_generation_server/models/ct2_causal_lm.py

@@ -63,7 +63,7 @@ class CT2CausalLM(Model):
         if ctranslate2 is None:
             raise ValueError(
                 "for quantization with ct2, the installation requires the pip package ctranslate2. "
-                "install via `text-generation-server[ct2]` or pip install ctranslate2 is required.",
+                "install via `text-generation-server[ct2]` or `pip install ctranslate2` is required.",
             )
 
         tokenizer = AutoTokenizer.from_pretrained(