Update README.md

README.md

@@ -185,119 +185,69 @@ extra_gated_fields:
   By clicking Submit below I accept the terms of the license and acknowledge that the information I provide will be collected stored processed and shared in accordance with the Meta Privacy Policy: checkbox
 extra_gated_description: The information you provide will be collected, stored, processed and shared in accordance with the [Meta Privacy Policy](https://www.facebook.com/privacy/policy/).
 extra_gated_button_content: Submit
-widget:
-  - example_title: Winter holidays
-    messages:
-    - role: system
-      content: You are a helpful and honest assistant. Please, respond concisely and truthfully.
-    - role: user
-      content: Can you recommend a good destination for Winter holidays?
-  - example_title: Programming assistant
-    messages:
-    - role: system
-      content: You are a helpful and honest code and programming assistant. Please, respond concisely and truthfully.
-    - role: user
-      content: Write a function that computes the nth fibonacci number.
-inference:
-  parameters:
-    max_new_tokens: 300
-    stop:
-    - <|end_of_text|>
-    - <|eot_id|>
 quantized_by: bartowski
+lm_studio:
+  param_count: 70b
+  use_case: general
+  release_date: 18-04-2024
+  model_creator: meta-llama
+  prompt_template: Llama 3
+  system_prompt: You are a helpful AI assistant.
+  base_model: llama
+  original_repo: meta-llama/Meta-Llama-3-70B-Instruct
 ---
 
-## Llamacpp imatrix Quantizations of Meta-Llama-3-70B-Instruct
+## 💫 Community Model> Llama 3 70B Instruct by Meta
 
-Using <a href="https://github.com/ggerganov/llama.cpp/">llama.cpp</a> release <a href="https://github.com/ggerganov/llama.cpp/releases/tag/b2777">b2777</a> for quantization.
+*👾 [LM Studio](https://lmstudio.ai) Community models highlights program. Highlighting new & noteworthy models by the community. Join the conversation on [Discord](https://discord.gg/aPQfnNkxGC)*.
 
-Original model: https://huggingface.co/meta-llama/Meta-Llama-3-70B-Instruct
+**Model creator:** [meta-llama](https://huggingface.co/meta-llama)<br>
+**Original model**: [Meta-Llama-3-70B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-70B-Instruct)<br>
+**GGUF quantization:** provided by [bartowski](https://huggingface.co/bartowski) based on `llama.cpp` release [b2777](https://github.com/ggerganov/llama.cpp/releases/tag/b2777)<br>
 
-All quants made using imatrix option with dataset provided by Kalomaze [here](https://github.com/ggerganov/llama.cpp/discussions/5263#discussioncomment-8395384)
+## Model Summary:
+Llama 3 represents a huge update to the Llama family of models. This model is the 70B parameter instruction tuned model, with performance reaching and usually exceeding GPT-3.5.<br>
+This is a massive milestone, as an open model reaches the performance of a closed model over double its size.<br>
+This model is very happy to follow the given system prompt, so use this to your advantage to get the behavior you desire.<br>
+Llama 3 excels at all the general usage situations, including multi turn conversations, general world knowledge, and coding.<br>
 
-## Prompt format
+This model is made with the BPE fixes from llama.cpp
 
-```
-<|begin_of_text|><|start_header_id|>system<|end_header_id|>
+## Prompt Template:
 
-{system_prompt}<|eot_id|><|start_header_id|>user<|end_header_id|>
-
-{prompt}<|eot_id|><|start_header_id|>assistant<|end_header_id|>
+Choose the 'Llama 3' preset in your LM Studio. 
 
+Under the hood, the model will see a prompt that's formatted like so:
 
 ```
+<|begin_of_text|><|start_header_id|>system<|end_header_id|>
 
-## Download a file (not the whole branch) from below:
-
-| Filename | Quant type | File Size | Description |
-| -------- | ---------- | --------- | ----------- |
-| [Meta-Llama-3-70B-Instruct-Q8_0.gguf](https://huggingface.co/bartowski/Meta-Llama-3-70B-Instruct-GGUF/tree/main/Meta-Llama-3-70B-Instruct-Q8_0.gguf) | Q8_0 | 74.97GB | Extremely high quality, generally unneeded but max available quant. |
-| [Meta-Llama-3-70B-Instruct-Q6_K.gguf](https://huggingface.co/bartowski/Meta-Llama-3-70B-Instruct-GGUF/tree/main/Meta-Llama-3-70B-Instruct-Q6_K.gguf) | Q6_K | 57.88GB | Very high quality, near perfect, *recommended*. |
-| [Meta-Llama-3-70B-Instruct-Q5_K_M.gguf](https://huggingface.co/bartowski/Meta-Llama-3-70B-Instruct-GGUF/blob/main/Meta-Llama-3-70B-Instruct-Q5_K_M.gguf) | Q5_K_M | 49.94GB | High quality, *recommended*. |
-| [Meta-Llama-3-70B-Instruct-Q5_K_S.gguf](https://huggingface.co/bartowski/Meta-Llama-3-70B-Instruct-GGUF/blob/main/Meta-Llama-3-70B-Instruct-Q5_K_S.gguf) | Q5_K_S | 48.65GB | High quality, *recommended*. |
-| [Meta-Llama-3-70B-Instruct-Q4_K_M.gguf](https://huggingface.co/bartowski/Meta-Llama-3-70B-Instruct-GGUF/blob/main/Meta-Llama-3-70B-Instruct-Q4_K_M.gguf) | Q4_K_M | 42.52GB | Good quality, uses about 4.83 bits per weight, *recommended*. |
-| [Meta-Llama-3-70B-Instruct-Q4_K_S.gguf](https://huggingface.co/bartowski/Meta-Llama-3-70B-Instruct-GGUF/blob/main/Meta-Llama-3-70B-Instruct-Q4_K_S.gguf) | Q4_K_S | 40.34GB | Slightly lower quality with more space savings, *recommended*. |
-| [Meta-Llama-3-70B-Instruct-IQ4_NL.gguf](https://huggingface.co/bartowski/Meta-Llama-3-70B-Instruct-GGUF/blob/main/Meta-Llama-3-70B-Instruct-IQ4_NL.gguf) | IQ4_NL | 40.05GB | Decent quality, slightly smaller than Q4_K_S with similar performance *recommended*. |
-| [Meta-Llama-3-70B-Instruct-IQ4_XS.gguf](https://huggingface.co/bartowski/Meta-Llama-3-70B-Instruct-GGUF/blob/main/Meta-Llama-3-70B-Instruct-IQ4_XS.gguf) | IQ4_XS | 37.90GB | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. |
-| [Meta-Llama-3-70B-Instruct-Q3_K_L.gguf](https://huggingface.co/bartowski/Meta-Llama-3-70B-Instruct-GGUF/blob/main/Meta-Llama-3-70B-Instruct-Q3_K_L.gguf) | Q3_K_L | 37.14GB | Lower quality but usable, good for low RAM availability. |
-| [Meta-Llama-3-70B-Instruct-Q3_K_M.gguf](https://huggingface.co/bartowski/Meta-Llama-3-70B-Instruct-GGUF/blob/main/Meta-Llama-3-70B-Instruct-Q3_K_M.gguf) | Q3_K_M | 34.26GB | Even lower quality. |
-| [Meta-Llama-3-70B-Instruct-IQ3_M.gguf](https://huggingface.co/bartowski/Meta-Llama-3-70B-Instruct-GGUF/blob/main/Meta-Llama-3-70B-Instruct-IQ3_M.gguf) | IQ3_M | 31.93GB | Medium-low quality, new method with decent performance comparable to Q3_K_M. |
-| [Meta-Llama-3-70B-Instruct-IQ3_S.gguf](https://huggingface.co/bartowski/Meta-Llama-3-70B-Instruct-GGUF/blob/main/Meta-Llama-3-70B-Instruct-IQ3_S.gguf) | IQ3_S | 30.91GB | Lower quality, new method with decent performance, recommended over Q3_K_S quant, same size with better performance. |
-| [Meta-Llama-3-70B-Instruct-Q3_K_S.gguf](https://huggingface.co/bartowski/Meta-Llama-3-70B-Instruct-GGUF/blob/main/Meta-Llama-3-70B-Instruct-Q3_K_S.gguf) | Q3_K_S | 30.91GB | Low quality, not recommended. |
-| [Meta-Llama-3-70B-Instruct-IQ3_XS.gguf](https://huggingface.co/bartowski/Meta-Llama-3-70B-Instruct-GGUF/blob/main/Meta-Llama-3-70B-Instruct-IQ3_XS.gguf) | IQ3_XS | 29.30GB | Lower quality, new method with decent performance, slightly better than Q3_K_S. |
-| [Meta-Llama-3-70B-Instruct-IQ3_XXS.gguf](https://huggingface.co/bartowski/Meta-Llama-3-70B-Instruct-GGUF/blob/main/Meta-Llama-3-70B-Instruct-IQ3_XXS.gguf) | IQ3_XXS | 27.46GB | Lower quality, new method with decent performance, comparable to Q3 quants. |
-| [Meta-Llama-3-70B-Instruct-Q2_K.gguf](https://huggingface.co/bartowski/Meta-Llama-3-70B-Instruct-GGUF/blob/main/Meta-Llama-3-70B-Instruct-Q2_K.gguf) | Q2_K | 26.37GB | Very low quality but surprisingly usable. |
-| [Meta-Llama-3-70B-Instruct-IQ2_M.gguf](https://huggingface.co/bartowski/Meta-Llama-3-70B-Instruct-GGUF/blob/main/Meta-Llama-3-70B-Instruct-IQ2_M.gguf) | IQ2_M | 24.11GB | Very low quality, uses SOTA techniques to also be surprisingly usable. |
-| [Meta-Llama-3-70B-Instruct-IQ2_S.gguf](https://huggingface.co/bartowski/Meta-Llama-3-70B-Instruct-GGUF/blob/main/Meta-Llama-3-70B-Instruct-IQ2_S.gguf) | IQ2_S | 22.24GB | Very low quality, uses SOTA techniques to be usable. |
-| [Meta-Llama-3-70B-Instruct-IQ2_XS.gguf](https://huggingface.co/bartowski/Meta-Llama-3-70B-Instruct-GGUF/blob/main/Meta-Llama-3-70B-Instruct-IQ2_XS.gguf) | IQ2_XS | 21.14GB | Very low quality, uses SOTA techniques to be usable. |
-| [Meta-Llama-3-70B-Instruct-IQ2_XXS.gguf](https://huggingface.co/bartowski/Meta-Llama-3-70B-Instruct-GGUF/blob/main/Meta-Llama-3-70B-Instruct-IQ2_XXS.gguf) | IQ2_XXS | 19.09GB | Lower quality, uses SOTA techniques to be usable. |
-| [Meta-Llama-3-70B-Instruct-IQ1_M.gguf](https://huggingface.co/bartowski/Meta-Llama-3-70B-Instruct-GGUF/blob/main/Meta-Llama-3-70B-Instruct-IQ1_M.gguf) | IQ1_M | 16.75GB | Extremely low quality, *not* recommended. |
-| [Meta-Llama-3-70B-Instruct-IQ1_S.gguf](https://huggingface.co/bartowski/Meta-Llama-3-70B-Instruct-GGUF/blob/main/Meta-Llama-3-70B-Instruct-IQ1_S.gguf) | IQ1_S | 15.34GB | Extremely low quality, *not* recommended. |
-
-## Downloading using huggingface-cli
-
-First, make sure you have hugginface-cli installed:
-
-```
-pip install -U "huggingface_hub[cli]"
-```
-
-Then, you can target the specific file you want:
-
-```
-huggingface-cli download bartowski/Meta-Llama-3-70B-Instruct-GGUF --include "Meta-Llama-3-70B-Instruct-Q4_K_M.gguf" --local-dir ./ --local-dir-use-symlinks False
-```
+{system_prompt}<|eot_id|><|start_header_id|>user<|end_header_id|>
 
-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:
+{prompt}<|eot_id|><|start_header_id|>assistant<|end_header_id|>
 
 ```
-huggingface-cli download bartowski/Meta-Llama-3-70B-Instruct-GGUF --include "Meta-Llama-3-70B-Instruct-Q8_0.gguf/*" --local-dir Meta-Llama-3-70B-Instruct-Q8_0 --local-dir-use-symlinks False
-```
-
-You can either specify a new local-dir (Meta-Llama-3-70B-Instruct-Q8_0) or download them all in place (./)
-
-## Which file should I choose?
 
-A great write up with charts showing various performances is provided by Artefact2 [here](https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9)
+Use cases and examples to come.
 
-The first thing to figure out is how big a model you can run. To do this, you'll need to figure out how much RAM and/or VRAM you have.
+## Technical Details
 
-If you want your model running as FAST as possible, you'll want to fit the whole thing on your GPU's VRAM. Aim for a quant with a file size 1-2GB smaller than your GPU's total VRAM.
+Llama 3 was trained on over 15T tokens from a massively diverse range of subjects and languages, and includes 4 times more code than Llama 2.
 
-If you want the absolute maximum quality, add both your system RAM and your GPU's VRAM together, then similarly grab a quant with a file size 1-2GB Smaller than that total.
+This model also features Grouped Attention Query (GQA) so that memory usage scales nicely over large contexts.
 
-Next, you'll need to decide if you want to use an 'I-quant' or a 'K-quant'.
+Instruction fine tuning was performed with a combination of supervised fine-tuning (SFT), rejection sampling, proximal policy optimization (PPO), and direct policy optimization (DPO).
 
-If you don't want to think too much, grab one of the K-quants. These are in format 'QX_K_X', like Q5_K_M.
+Only IQ1_M and IQ2_XS use importance matrix (iMatrix), the rest are made with the standard quant algorithms.
 
-If you want to get more into the weeds, you can check out this extremely useful feature chart:
+Check out their blog post for more information [here](https://ai.meta.com/blog/meta-llama-3/)
 
-[llama.cpp feature matrix](https://github.com/ggerganov/llama.cpp/wiki/Feature-matrix)
+## Special thanks
 
-But basically, if you're aiming for below Q4, and you're running cuBLAS (Nvidia) or rocBLAS (AMD), you should look towards the I-quants. These are in format IQX_X, like IQ3_M. These are newer and offer better performance for their size.
+🙏 Special thanks to [Georgi Gerganov](https://github.com/ggerganov) and the whole team working on [llama.cpp](https://github.com/ggerganov/llama.cpp/) for making all of this possible.
 
-These I-quants can also be used on CPU and Apple Metal, but will be slower than their K-quant equivalent, so speed vs performance is a tradeoff you'll have to decide.
+🙏 Special thanks to [Kalomaze](https://github.com/kalomaze) for his dataset (linked [here](https://github.com/ggerganov/llama.cpp/discussions/5263)) that was used for calculating the imatrix for the IQ1_M and IQ2_XS quants, which makes them usable even at their tiny size!
 
-The I-quants are *not* compatible with Vulcan, which is also AMD, so if you have an AMD card double check if you're using the rocBLAS build or the Vulcan build. At the time of writing this, LM Studio has a preview with ROCm support, and other inference engines have specific builds for ROCm.
+## Disclaimers
 
-Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski
+LM Studio is not the creator, originator, or owner of any Model featured in the Community Model Program. Each Community Model is created and provided by third parties. LM Studio does not endorse, support, represent or guarantee the completeness, truthfulness, accuracy, or reliability of any Community Model.  You understand that Community Models can produce content that might be offensive, harmful, inaccurate or otherwise inappropriate, or deceptive. Each Community Model is the sole responsibility of the person or entity who originated such Model. LM Studio may not monitor or control the Community Models and cannot, and does not, take responsibility for any such Model. LM Studio disclaims all warranties or guarantees about the accuracy, reliability or benefits of the Community Models.  LM Studio further disclaims any warranty that the Community Model will meet your requirements, be secure, uninterrupted or available at any time or location, or error-free, viruses-free, or that any errors will be corrected, or otherwise. You will be solely responsible for any damage resulting from your use of or access to the Community Models, your downloading of any Community Model, or use of any other Community Model provided by or through LM Studio.