Spaces:

yasu-oh
/

quantized_model_memory_estimator

Sleeping

yasu-oh commited on May 22

Commit

06ceb45

1 Parent(s): 5df4a84

IQ4_XS

Files changed (1) hide show

app.py CHANGED Viewed

@@ -8,7 +8,7 @@ from typing import Optional
 API_BASE_URL = "https://huggingface.co/api/models/"
 LLAMA_CPP_SOURCE = "https://raw.githubusercontent.com/ggml-org/llama.cpp/refs/heads/master/convert_hf_to_gguf.py"
 QUANTIZE = {
-    "IQ4_XS": 4.25,
     "Q4_K_M": 4.9,
     "Q5_K_M": 5.7,
     "Q6_K": 6.6,

 API_BASE_URL = "https://huggingface.co/api/models/"
 LLAMA_CPP_SOURCE = "https://raw.githubusercontent.com/ggml-org/llama.cpp/refs/heads/master/convert_hf_to_gguf.py"
 QUANTIZE = {
+    "IQ4_XS": 4.3,
     "Q4_K_M": 4.9,
     "Q5_K_M": 5.7,
     "Q6_K": 6.6,