update readme table

README.md

@@ -22,17 +22,19 @@ Starting out with Q4_K_M, and iterating from there. Future plans for imatrix/IQ
 
 First time doing quantizations so any feedback is greatly appreciated.
 
-| Name | Quant method | Bits |
-| ---- | ---- | ---- |
-| [Mixtral-Instruct-ITR-8x7B.Q2_K.gguf](https://huggingface.co/InferenceIllusionist/Mixtral-Instruct-ITR-8x7B-GGUF/blob/main/Mixtral-Instruct-ITR-8x7B-Q2_K.gguf) | Q2_K | 2 |
-| [Mixtral-Instruct-ITR-8x7B.Q3_K_S.gguf](https://huggingface.co/InferenceIllusionist/Mixtral-Instruct-ITR-8x7B-GGUF/blob/main/Mixtral-Instruct-ITR-8x7B-Q3_K_S.gguf) | Q3_K_S | 3 |
-| [Mixtral-Instruct-ITR-8x7B.Q3_K_M.gguf](https://huggingface.co/InferenceIllusionist/Mixtral-Instruct-ITR-8x7B-GGUF/blob/main/Mixtral-Instruct-ITR-8x7B-Q3_K_M.gguf) | Q3_K_M | 3 |
-| [Mixtral-Instruct-ITR-8x7B.Q3_K_L.gguf](https://huggingface.co/InferenceIllusionist/Mixtral-Instruct-ITR-8x7B-GGUF/blob/main/Mixtral-Instruct-ITR-8x7B-Q3_K_L.gguf) | Q3_K_L | 4 |
-| [Mixtral-Instruct-ITR-8x7B.Q4_K_M.gguf](https://huggingface.co/InferenceIllusionist/Mixtral-Instruct-ITR-8x7B-GGUF/blob/main/Mixtral-Instruct-ITR-8x7B-Q4_K_M.gguf) | Q4_K_M | 5 | 
-| [Mixtral-Instruct-ITR-8x7B.Q5_K_S.gguf](https://huggingface.co/InferenceIllusionist/Mixtral-Instruct-ITR-8x7B-GGUF/blob/main/Mixtral-Instruct-ITR-8x7B-Q5_K_S.gguf) | Q5_K_S | 5 | 
-| [Mixtral-Instruct-ITR-8x7B.Q5_K_M.gguf](https://huggingface.co/InferenceIllusionist/Mixtral-Instruct-ITR-8x7B-GGUF/blob/main/Mixtral-Instruct-ITR-8x7B-Q5_K_M.gguf) | Q6_K_M | 5 | 
-| [Mixtral-Instruct-ITR-8x7B.Q6_K.gguf](https://huggingface.co/InferenceIllusionist/Mixtral-Instruct-ITR-8x7B-GGUF/blob/main/Mixtral-Instruct-ITR-8x7B-Q6_K.gguf) | Q6_K | 6 | 49.62 GB| 
-| [Mixtral-Instruct-ITR-8x7B.Q8_0.gguf](https://huggingface.co/InferenceIllusionist/Mixtral-Instruct-ITR-8x7B-GGUF/blob/main/Mixtral-Instruct-ITR-8x7B-Q8_0.gguf) | Q8_0 | 8 | 49.62 GB| 
+| Name | Quant method | Bits |ppl*
+| ---- | ---- | ---- | ---- |
+| [Mixtral-Instruct-ITR-8x7B.Q2_K.gguf](https://huggingface.co/InferenceIllusionist/Mixtral-Instruct-ITR-8x7B-GGUF/blob/main/Mixtral-Instruct-ITR-8x7B-Q2_K.gguf) | Q2_K | 2 | +0.6717 ppl|
+| [Mixtral-Instruct-ITR-8x7B.Q3_K_S.gguf](https://huggingface.co/InferenceIllusionist/Mixtral-Instruct-ITR-8x7B-GGUF/blob/main/Mixtral-Instruct-ITR-8x7B-Q3_K_S.gguf) | Q3_K_S | 3 | +0.5551 ppl|
+| [Mixtral-Instruct-ITR-8x7B.Q3_K_M.gguf](https://huggingface.co/InferenceIllusionist/Mixtral-Instruct-ITR-8x7B-GGUF/blob/main/Mixtral-Instruct-ITR-8x7B-Q3_K_M.gguf) | Q3_K_M | 3 | +0.2496 ppl|
+| [Mixtral-Instruct-ITR-8x7B.Q3_K_L.gguf](https://huggingface.co/InferenceIllusionist/Mixtral-Instruct-ITR-8x7B-GGUF/blob/main/Mixtral-Instruct-ITR-8x7B-Q3_K_L.gguf) | Q3_K_L | 4 | +0.1764 ppl|
+| [Mixtral-Instruct-ITR-8x7B.Q4_K_M.gguf](https://huggingface.co/InferenceIllusionist/Mixtral-Instruct-ITR-8x7B-GGUF/blob/main/Mixtral-Instruct-ITR-8x7B-Q4_K_M.gguf) | Q4_K_M | 5 | +0.0532 ppl|
+| [Mixtral-Instruct-ITR-8x7B.Q5_K_S.gguf](https://huggingface.co/InferenceIllusionist/Mixtral-Instruct-ITR-8x7B-GGUF/blob/main/Mixtral-Instruct-ITR-8x7B-Q5_K_S.gguf) | Q5_K_S | 5 | +0.0400 ppl|
+| [Mixtral-Instruct-ITR-8x7B.Q5_K_M.gguf](https://huggingface.co/InferenceIllusionist/Mixtral-Instruct-ITR-8x7B-GGUF/blob/main/Mixtral-Instruct-ITR-8x7B-Q5_K_M.gguf) | Q5_K_M | 6 | +0.0122 ppl|
+| [Mixtral-Instruct-ITR-8x7B.Q6_K.gguf](https://huggingface.co/InferenceIllusionist/Mixtral-Instruct-ITR-8x7B-GGUF/blob/main/Mixtral-Instruct-ITR-8x7B-Q6_K.gguf) | Q6_K | 6 | +0.008 ppl|  
+| [Mixtral-Instruct-ITR-8x7B.Q8_0.gguf](https://huggingface.co/InferenceIllusionist/Mixtral-Instruct-ITR-8x7B-GGUF/blob/main/Mixtral-Instruct-ITR-8x7B-Q8_0.gguf) | Q8_0 | 8 | +0.004 ppl|
+
+*Perplexity @ LLaMA-v1-7B for reference
 
 Original model card below for reference.