Upload README.md with huggingface_hub

README.md

@@ -1,28 +1,11 @@
 ---
-base_model: mistralai/Mistral-Large-Instruct-2407
-language:
-- en
-- fr
-- de
-- es
-- it
-- pt
-- zh
-- ja
-- ru
-- ko
-license: other
-license_name: mrl
-license_link: https://mistral.ai/licenses/MRL-0.1.md
-pipeline_tag: text-generation
 quantized_by: bartowski
-extra_gated_description: If you want to learn more about how we process your personal
-  data, please read our <a href="https://mistral.ai/terms/">Privacy Policy</a>.
+pipeline_tag: text-generation
 ---
 
 ## Llamacpp imatrix Quantizations of Mistral-Large-Instruct-2407
 
-Using <a href="https://github.com/ggerganov/llama.cpp/">llama.cpp</a> release <a href="https://github.com/ggerganov/llama.cpp/releases/tag/b3460">b3460</a> for quantization.
+Using <a href="https://github.com/ggerganov/llama.cpp/">llama.cpp</a> release <a href="https://github.com/ggerganov/llama.cpp/releases/tag/b3634">b3634</a> for quantization.
 
 Original model: https://huggingface.co/mistralai/Mistral-Large-Instruct-2407
 
@@ -33,9 +16,13 @@ Run them in [LM Studio](https://lmstudio.ai/)
 ## Prompt format
 
 ```
-<s>[INST] {prompt}[/INST] </s>
+<s>[INST] {prompt}[/INST] 
 ```
 
+## What's new:
+
+Add chat template, some new sizes
+
 ## Download a file (not the whole branch) from below:
 
 | Filename | Quant type | File Size | Split | Description |
@@ -43,10 +30,11 @@ Run them in [LM Studio](https://lmstudio.ai/)
 | [Mistral-Large-Instruct-2407-Q8_0.gguf](https://huggingface.co/bartowski/Mistral-Large-Instruct-2407-GGUF/tree/main/Mistral-Large-Instruct-2407-Q8_0) | Q8_0 | 130.28GB | true | Extremely high quality, generally unneeded but max available quant. |
 | [Mistral-Large-Instruct-2407-Q6_K.gguf](https://huggingface.co/bartowski/Mistral-Large-Instruct-2407-GGUF/tree/main/Mistral-Large-Instruct-2407-Q6_K) | Q6_K | 100.59GB | true | Very high quality, near perfect, *recommended*. |
 | [Mistral-Large-Instruct-2407-Q5_K_M.gguf](https://huggingface.co/bartowski/Mistral-Large-Instruct-2407-GGUF/tree/main/Mistral-Large-Instruct-2407-Q5_K_M) | Q5_K_M | 86.49GB | true | High quality, *recommended*. |
-| [Mistral-Large-Instruct-2407-Q5_K_S.gguf](https://huggingface.co/bartowski/Mistral-Large-Instruct-2407-GGUF/tree/main/Mistral-Large-Instruct-2407-Q5_K_S) | Q5_K_S | 84.36GB | true | High quality, *recommended*. |
 | [Mistral-Large-Instruct-2407-Q4_K_M.gguf](https://huggingface.co/bartowski/Mistral-Large-Instruct-2407-GGUF/tree/main/Mistral-Large-Instruct-2407-Q4_K_M) | Q4_K_M | 73.22GB | true | Good quality, default size for must use cases, *recommended*. |
-| [Mistral-Large-Instruct-2407-IQ4_XS.gguf](https://huggingface.co/bartowski/Mistral-Large-Instruct-2407-GGUF/tree/main/Mistral-Large-Instruct-2407-IQ4_XS) | IQ4_XS | 65.43GB | true | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. |
 | [Mistral-Large-Instruct-2407-Q4_K_S.gguf](https://huggingface.co/bartowski/Mistral-Large-Instruct-2407-GGUF/tree/main/Mistral-Large-Instruct-2407-Q4_K_S) | Q4_K_S | 69.57GB | true | Slightly lower quality with more space savings, *recommended*. |
+| [Mistral-Large-Instruct-2407-Q4_0.gguf](https://huggingface.co/bartowski/Mistral-Large-Instruct-2407-GGUF/tree/main/Mistral-Large-Instruct-2407-Q4_0) | Q4_0 | 69.32GB | true | Legacy format, generally not worth using over similarly sized formats |
+| [Mistral-Large-Instruct-2407-Q4_0_4_4.gguf](https://huggingface.co/bartowski/Mistral-Large-Instruct-2407-GGUF/tree/main/Mistral-Large-Instruct-2407-Q4_0_4_4) | Q4_0_4_4 | 69.08GB | true | Optimized for ARM and CPU inference, much faster than Q4_0 at similar quality. |
+| [Mistral-Large-Instruct-2407-IQ4_XS.gguf](https://huggingface.co/bartowski/Mistral-Large-Instruct-2407-GGUF/tree/main/Mistral-Large-Instruct-2407-IQ4_XS) | IQ4_XS | 65.43GB | true | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. |
 | [Mistral-Large-Instruct-2407-Q3_K_XL.gguf](https://huggingface.co/bartowski/Mistral-Large-Instruct-2407-GGUF/tree/main/Mistral-Large-Instruct-2407-Q3_K_XL) | Q3_K_XL | 64.91GB | true | Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability. |
 | [Mistral-Large-Instruct-2407-Q3_K_L.gguf](https://huggingface.co/bartowski/Mistral-Large-Instruct-2407-GGUF/tree/main/Mistral-Large-Instruct-2407-Q3_K_L) | Q3_K_L | 64.55GB | true | Lower quality but usable, good for low RAM availability. |
 | [Mistral-Large-Instruct-2407-Q3_K_M.gguf](https://huggingface.co/bartowski/Mistral-Large-Instruct-2407-GGUF/tree/main/Mistral-Large-Instruct-2407-Q3_K_M) | Q3_K_M | 59.10GB | true | Low quality. |
@@ -60,6 +48,14 @@ Run them in [LM Studio](https://lmstudio.ai/)
 | [Mistral-Large-Instruct-2407-IQ2_XXS.gguf](https://huggingface.co/bartowski/Mistral-Large-Instruct-2407-GGUF/blob/main/Mistral-Large-Instruct-2407-IQ2_XXS.gguf) | IQ2_XXS | 32.43GB | false | Very low quality, uses SOTA techniques to be usable. |
 | [Mistral-Large-Instruct-2407-IQ1_M.gguf](https://huggingface.co/bartowski/Mistral-Large-Instruct-2407-GGUF/blob/main/Mistral-Large-Instruct-2407-IQ1_M.gguf) | IQ1_M | 28.39GB | false | Extremely low quality, *not* recommended. |
 
+## Embed/output weights
+
+Some of these quants (Q3_K_XL, Q4_K_L etc) are the standard quantization method with the embeddings and output weights quantized to Q8_0 instead of what they would normally default to.
+
+Some say that this improves the quality, others don't notice any difference. If you use these models PLEASE COMMENT with your findings. I would like feedback that these are actually used and useful so I don't keep uploading quants no one is using.
+
+Thanks!
+
 ## Credits
 
 Thank you kalomaze and Dampf for assistance in creating the imatrix calibration dataset
@@ -83,7 +79,7 @@ huggingface-cli download bartowski/Mistral-Large-Instruct-2407-GGUF --include "M
 If the model is bigger than 50GB, it will have been split into multiple files. In order to download them all to a local folder, run:
 
 ```
-huggingface-cli download bartowski/Mistral-Large-Instruct-2407-GGUF --include "Mistral-Large-Instruct-2407-Q8_0.gguf/*" --local-dir Mistral-Large-Instruct-2407-Q8_0
+huggingface-cli download bartowski/Mistral-Large-Instruct-2407-GGUF --include "Mistral-Large-Instruct-2407-Q8_0/*" --local-dir ./
 ```
 
 You can either specify a new local-dir (Mistral-Large-Instruct-2407-Q8_0) or download them all in place (./)