Update LLM wirh REPO_ID

.ruff.toml

@@ -0,0 +1,4 @@
+# Assume Python 3.10.
+target-version = "py310"
+# Decrease the maximum line length to 79 characters.
+line-length = 300

.stignore

@@ -1,3 +1,4 @@
+models
 .git
 # Byte-compiled / optimized / DLL files
 __pycache__

app.py

@@ -6,9 +6,9 @@
 # gradio.load("models/WizardLM/WizardCoder-15B-V1.0").launch()
 
 import os
-from pathlib import Path
 import time
 from dataclasses import asdict, dataclass
+from pathlib import Path
 from types import SimpleNamespace
 
 import gradio as gr
@@ -193,7 +193,7 @@ MODEL_FILENAME = "WizardCoder-15B-1.0.ggmlv3.q4_1.bin"  # 11.9G
 MODEL_FILENAME = "WizardCoder-15B-1.0.ggmlv3.q4_1.bin"  # 11.9G
 
 # https://huggingface.co/TheBloke/WizardLM-13B-V1.0-Uncensored-GGML
-MODEL_FILENAME = "wizardlm-13b-v1.0-uncensored.ggmlv3.q4_1.bin"
+MODEL_FILENAME = "wizardlm-13b-v1.0-uncensored.ggmlv3.q4_1.bin"  # 8.4G
 
 DESTINATION_FOLDER = "models"
 
@@ -204,8 +204,8 @@ if "WizardCoder" in MODEL_FILENAME:
 if "uncensored" in MODEL_FILENAME.lower():
     REPO_ID = "TheBloke/WizardLM-13B-V1.0-Uncensored-GGML"
 
+logger.info("start dl")
 download_quant(DESTINATION_FOLDER, REPO_ID, MODEL_FILENAME)
-
 logger.info("done dl")
 
 # if "mpt" in model_filename:
@@ -227,21 +227,18 @@ llm = AutoModelForCausalLM.from_pretrained(
 # """
 
 logger.debug(f"{os.cpu_count()=}")
+logger.info("load llm")
 
-if "WizardCoder" in MODEL_FILENAME:
-    _ = Path("models", MODEL_FILENAME).absolute().as_posix()
-    LLM = AutoModelForCausalLM.from_pretrained(
-        "TheBloke/WizardCoder-15B-1.0-GGML",
-        model_file=_,
-        model_type="starcoder",
-        threads=os.cpu_count() // 2,  # type: ignore
-    )
-# LLM = AutoModelForCausalLM.from_pretrained(
-# "TheBloke/WizardCoder-15B-1.0-GGML",
-# model_file=MODEL_FILENAME,
-# model_type="starcoder",
-# threads=os.cpu_count() // 2  # type: ignore
-# )
+_ = Path("models", MODEL_FILENAME).absolute().as_posix()
+LLM = AutoModelForCausalLM.from_pretrained(
+    # "TheBloke/WizardCoder-15B-1.0-GGML",
+    REPO_ID,
+    model_file=_,
+    model_type="starcoder",
+    threads=os.cpu_count() // 2,  # type: ignore
+)
+
+logger.info("done load llm")
 
 cpu_count = os.cpu_count() // 2  # type: ignore
 logger.debug(f"{cpu_count=}")