Fix typo in js script, add printing to console for preview node, remove deprecated config option

README.md

@@ -43,11 +43,6 @@ git clone https://huggingface.co/turboderp/Llama-3.1-8B-Instruct-exl2 -b 4.0bpw
     <td><i>cache_bits</i></td>
     <td>A lower value reduces VRAM usage, but also affects generation speed and quality.</td>
   </tr>
-  <tr>
-    <td></td>
-    <td><i>fast_tensors</i></td>
-    <td>Enabling reduces RAM usage and speeds up model loading.</td>
-  </tr>
   <tr>
     <td></td>
     <td ><i>flash_attention</i></td>

exllama.py

@@ -27,12 +27,11 @@ from folder_paths import add_model_folder_path, get_folder_paths, models_dir
 _CATEGORY = "zuellni/exllama"
 _MAPPING = "ZuellniExLlama"
 
-class Loader:
-    _input_info = None
 
+class Loader:
     @classmethod
     def INPUT_TYPES(cls):
-        def get_input_info(cls):
+        if not cls._MODELS:
             add_model_folder_path("llm", str(Path(models_dir) / "llm"))
 
             for folder in get_folder_paths("llm"):
@@ -45,17 +44,10 @@ class Loader:
         caches = list(cls._CACHES.keys())
         default = models[0] if models else None
 
-            return models, caches, default
-
-        if Loader._input_info is None:
-            Loader._input_info = get_input_info(cls)
-        models, caches, default = Loader._input_info
-
         return {
             "required": {
                 "model": (models, {"default": default}),
                 "cache_bits": (caches, {"default": 4}),
-                "fast_tensors": ("BOOLEAN", {"default": True}),
                 "flash_attention": ("BOOLEAN", {"default": True}),
                 "max_seq_len": (
                     "INT",
@@ -76,12 +68,11 @@ class Loader:
     RETURN_NAMES = ("MODEL",)
     RETURN_TYPES = ("EXL_MODEL",)
 
-    def setup(self, model, cache_bits, fast_tensors, flash_attention, max_seq_len):
+    def setup(self, model, cache_bits, flash_attention, max_seq_len):
         self.unload()
         self.cache_bits = cache_bits
 
         self.config = ExLlamaV2Config(__class__._MODELS[model])
-        self.config.fasttensors = fast_tensors
         self.config.no_flash_attn = not flash_attention
 
         if max_seq_len:
@@ -328,7 +319,7 @@ class Generator:
 
         if not settings:
             settings = ExLlamaV2Sampler.Settings()
-            settings.greedy()
+            settings = settings.greedy()
 
         job = ExLlamaV2DynamicJob(
             input_ids=tokens,

text.js

@@ -3,7 +3,7 @@ import { app } from "../../../scripts/app.js"
 app.registerExtension({
 	name: "ZuellniText",
 	async beforeRegisterNodeDef(nodeType, nodeData, app) {
-		if (nodeData.category != "Zuellni/Text")
+		if (nodeData.category != "zuellni/text")
 			return
 
 		const onNodeCreated = nodeType.prototype.onNodeCreated

text.py

@@ -71,6 +71,7 @@ class Preview:
         return {
             "required": {
                 "text": ("STRING", {"default": "", "forceInput": True}),
+                "print_to_console": ("BOOLEAN", {"default": False}),
                 "output": ("STRING", {"default": "", "multiline": True}),
             }
         }
@@ -80,7 +81,8 @@ class Preview:
     OUTPUT_NODE = True
     RETURN_TYPES = ()
 
-    def preview(self, text, output):
+    def preview(self, text, print_to_console, output):
+        print_to_console and print(text)
         return {"ui": {"text": [text]}}