Update README.md

README.md

@@ -83,17 +83,38 @@ model-index:
     - type: pass@1
       value: 40.6
 quantized_by: bartowski
+lm_studio:
+  param_count: 15b
+  use_case: coding
+  release_date: 30-04-2024
+  model_creator: BigCode
+  prompt_template: Starcoder2 Instruct
+  system_prompt: none
+  base_model: starcoder2
+  original_repo: bigcode/starcoder2-15b-instruct-v0.1
 ---
 
-## Llamacpp imatrix Quantizations of starcoder2-15b-instruct-v0.1
+## 💫 Community Model> Starcoder2 15B Instruct v0.1 by BigCode
 
-Using <a href="https://github.com/ggerganov/llama.cpp/">llama.cpp</a> release <a href="https://github.com/ggerganov/llama.cpp/releases/tag/b2756">b2756</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/bigcode/starcoder2-15b-instruct-v0.1
+**Model creator:** [bigcode](https://huggingface.co/bigcode)<br>
+**Original model**: [starcoder2-15b-instruct-v0.1](https://huggingface.co/bigcode/starcoder2-15b-instruct-v0.1)<br>
+**GGUF quantization:** provided by [bartowski](https://huggingface.co/bartowski) based on `llama.cpp` release [b2756](https://github.com/ggerganov/llama.cpp/releases/tag/b2756)<br>
 
-All quants made using imatrix option with dataset provided by Kalomaze [here](https://github.com/ggerganov/llama.cpp/discussions/5263#discussioncomment-8395384)
+We introduce StarCoder2-15B-Instruct-v0.1, the very first entirely self-aligned code Large Language Model (LLM) trained with a fully permissive and transparent pipeline. Our open-source pipeline uses StarCoder2-15B to generate thousands of instruction-response pairs, which are then used to fine-tune StarCoder-15B itself without any human annotations or distilled data from huge and proprietary LLMs.
 
-## Prompt format
+
+## Model Summary:
+Starcoder2-15B-Instruct-v0.1 is self-proclaimed to be the first entirely self-aligned code model with a fully permissive and transparent pipeline.<br>
+This model is meant to be used for coding instructions in a <b>single turn<b>, any other styles may result in less accurate responses.<br>
+Starcoder2 has been primarily finetuned for Python code generation and as such should primarily be used for Python tasks.
+
+## Prompt Template:
+
+Choose the 'Starcoder2 Instruct' preset in your LM Studio. 
+
+Under the hood, the model will see a prompt that's formatted like so:
 
 ```
 <|endoftext|>You are an exceptionally intelligent coding assistant that consistently delivers accurate and reliable responses to user instructions.
@@ -103,60 +124,41 @@ All quants made using imatrix option with dataset provided by Kalomaze [here](ht
 
 ### Response
 <|endoftext|>
-
 ```
 
-Note that this model does not support a System prompt.
+## Use case and examples
+
+This model should be used for single turn coding related instructions.
+
+## Coding with requirements
+
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/6435718aaaef013d1aec3b8b/rNqulMDumAp7s1LdIAerC.png)
 
-## Download a file (not the whole branch) from below:
+## Creating unit tests
 
-| Filename | Quant type | File Size | Description |
-| -------- | ---------- | --------- | ----------- |
-| [starcoder2-15b-instruct-v0.1-Q8_0.gguf](https://huggingface.co/bartowski/starcoder2-15b-instruct-v0.1-GGUF/blob/main/starcoder2-15b-instruct-v0.1-Q8_0.gguf) | Q8_0 | 16.96GB | Extremely high quality, generally unneeded but max available quant. |
-| [starcoder2-15b-instruct-v0.1-Q6_K.gguf](https://huggingface.co/bartowski/starcoder2-15b-instruct-v0.1-GGUF/blob/main/starcoder2-15b-instruct-v0.1-Q6_K.gguf) | Q6_K | 13.10GB | Very high quality, near perfect, *recommended*. |
-| [starcoder2-15b-instruct-v0.1-Q5_K_M.gguf](https://huggingface.co/bartowski/starcoder2-15b-instruct-v0.1-GGUF/blob/main/starcoder2-15b-instruct-v0.1-Q5_K_M.gguf) | Q5_K_M | 11.43GB | High quality, *recommended*. |
-| [starcoder2-15b-instruct-v0.1-Q5_K_S.gguf](https://huggingface.co/bartowski/starcoder2-15b-instruct-v0.1-GGUF/blob/main/starcoder2-15b-instruct-v0.1-Q5_K_S.gguf) | Q5_K_S | 11.02GB | High quality, *recommended*. |
-| [starcoder2-15b-instruct-v0.1-Q4_K_M.gguf](https://huggingface.co/bartowski/starcoder2-15b-instruct-v0.1-GGUF/blob/main/starcoder2-15b-instruct-v0.1-Q4_K_M.gguf) | Q4_K_M | 9.86GB | Good quality, uses about 4.83 bits per weight, *recommended*. |
-| [starcoder2-15b-instruct-v0.1-Q4_K_S.gguf](https://huggingface.co/bartowski/starcoder2-15b-instruct-v0.1-GGUF/blob/main/starcoder2-15b-instruct-v0.1-Q4_K_S.gguf) | Q4_K_S | 9.16GB | Slightly lower quality with more space savings, *recommended*. |
-| [starcoder2-15b-instruct-v0.1-IQ4_NL.gguf](https://huggingface.co/bartowski/starcoder2-15b-instruct-v0.1-GGUF/blob/main/starcoder2-15b-instruct-v0.1-IQ4_NL.gguf) | IQ4_NL | 9.08GB | Decent quality, slightly smaller than Q4_K_S with similar performance *recommended*. |
-| [starcoder2-15b-instruct-v0.1-IQ4_XS.gguf](https://huggingface.co/bartowski/starcoder2-15b-instruct-v0.1-GGUF/blob/main/starcoder2-15b-instruct-v0.1-IQ4_XS.gguf) | IQ4_XS | 8.59GB | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. |
-| [starcoder2-15b-instruct-v0.1-Q3_K_L.gguf](https://huggingface.co/bartowski/starcoder2-15b-instruct-v0.1-GGUF/blob/main/starcoder2-15b-instruct-v0.1-Q3_K_L.gguf) | Q3_K_L | 8.96GB | Lower quality but usable, good for low RAM availability. |
-| [starcoder2-15b-instruct-v0.1-Q3_K_M.gguf](https://huggingface.co/bartowski/starcoder2-15b-instruct-v0.1-GGUF/blob/main/starcoder2-15b-instruct-v0.1-Q3_K_M.gguf) | Q3_K_M | 8.04GB | Even lower quality. |
-| [starcoder2-15b-instruct-v0.1-IQ3_M.gguf](https://huggingface.co/bartowski/starcoder2-15b-instruct-v0.1-GGUF/blob/main/starcoder2-15b-instruct-v0.1-IQ3_M.gguf) | IQ3_M | 7.30GB | Medium-low quality, new method with decent performance comparable to Q3_K_M. |
-| [starcoder2-15b-instruct-v0.1-IQ3_S.gguf](https://huggingface.co/bartowski/starcoder2-15b-instruct-v0.1-GGUF/blob/main/starcoder2-15b-instruct-v0.1-IQ3_S.gguf) | IQ3_S | 7.00GB | Lower quality, new method with decent performance, recommended over Q3_K_S quant, same size with better performance. |
-| [starcoder2-15b-instruct-v0.1-Q3_K_S.gguf](https://huggingface.co/bartowski/starcoder2-15b-instruct-v0.1-GGUF/blob/main/starcoder2-15b-instruct-v0.1-Q3_K_S.gguf) | Q3_K_S | 6.98GB | Low quality, not recommended. |
-| [starcoder2-15b-instruct-v0.1-IQ3_XS.gguf](https://huggingface.co/bartowski/starcoder2-15b-instruct-v0.1-GGUF/blob/main/starcoder2-15b-instruct-v0.1-IQ3_XS.gguf) | IQ3_XS | 6.71GB | Lower quality, new method with decent performance, slightly better than Q3_K_S. |
-| [starcoder2-15b-instruct-v0.1-IQ3_XXS.gguf](https://huggingface.co/bartowski/starcoder2-15b-instruct-v0.1-GGUF/blob/main/starcoder2-15b-instruct-v0.1-IQ3_XXS.gguf) | IQ3_XXS | 6.21GB | Lower quality, new method with decent performance, comparable to Q3 quants. |
-| [starcoder2-15b-instruct-v0.1-Q2_K.gguf](https://huggingface.co/bartowski/starcoder2-15b-instruct-v0.1-GGUF/blob/main/starcoder2-15b-instruct-v0.1-Q2_K.gguf) | Q2_K | 6.19GB | Very low quality but surprisingly usable. |
-| [starcoder2-15b-instruct-v0.1-IQ2_M.gguf](https://huggingface.co/bartowski/starcoder2-15b-instruct-v0.1-GGUF/blob/main/starcoder2-15b-instruct-v0.1-IQ2_M.gguf) | IQ2_M | 5.54GB | Very low quality, uses SOTA techniques to also be surprisingly usable. |
-| [starcoder2-15b-instruct-v0.1-IQ2_S.gguf](https://huggingface.co/bartowski/starcoder2-15b-instruct-v0.1-GGUF/blob/main/starcoder2-15b-instruct-v0.1-IQ2_S.gguf) | IQ2_S | 5.14GB | Very low quality, uses SOTA techniques to be usable. |
-| [starcoder2-15b-instruct-v0.1-IQ2_XS.gguf](https://huggingface.co/bartowski/starcoder2-15b-instruct-v0.1-GGUF/blob/main/starcoder2-15b-instruct-v0.1-IQ2_XS.gguf) | IQ2_XS | 4.82GB | Very low quality, uses SOTA techniques to be usable. |
-| [starcoder2-15b-instruct-v0.1-IQ2_XXS.gguf](https://huggingface.co/bartowski/starcoder2-15b-instruct-v0.1-GGUF/blob/main/starcoder2-15b-instruct-v0.1-IQ2_XXS.gguf) | IQ2_XXS | 4.36GB | Lower quality, uses SOTA techniques to be usable. |
-| [starcoder2-15b-instruct-v0.1-IQ1_M.gguf](https://huggingface.co/bartowski/starcoder2-15b-instruct-v0.1-GGUF/blob/main/starcoder2-15b-instruct-v0.1-IQ1_M.gguf) | IQ1_M | 3.86GB | Extremely low quality, *not* recommended. |
-| [starcoder2-15b-instruct-v0.1-IQ1_S.gguf](https://huggingface.co/bartowski/starcoder2-15b-instruct-v0.1-GGUF/blob/main/starcoder2-15b-instruct-v0.1-IQ1_S.gguf) | IQ1_S | 3.55GB | Extremely low quality, *not* recommended. |
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/6435718aaaef013d1aec3b8b/q_VNUflz6tcAScY_yDLet.png)
 
-## Which file should I choose?
+## More coding examples
 
-A great write up with charts showing various performances is provided by Artefact2 [here](https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9)
 
-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.
+Starcoder2 15B instruct was trained primarily on Python code generation tasks. Using Starcoder2 15B (non instruct) to generated thousands of instruction-reponse pairs, the results were used to fine tune an instruct model without human annotation or distilled data.
 
-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.
+The dataset created is open and available: [self-oss-instruct-sc2-exec-filter-50k](https://huggingface.co/datasets/bigcode/self-oss-instruct-sc2-exec-filter-50k)
 
-Next, you'll need to decide if you want to use an 'I-quant' or a 'K-quant'.
+And the code used to create the self-alignment has been shared here: [starcoder2-self-align](https://github.com/bigcode-project/starcoder2-self-align)
 
-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.
+The results of the self-alignment are extremely promising, with significantly higher scores across all coding benchmarks, which is a great sign for future progress.
 
-If you want to get more into the weeds, you can check out this extremely useful feature chart:
+More details on their model card [here](https://huggingface.co/bigcode/starcoder2-15b-instruct-v0.1)
 
-[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 these quants, which improves the overall quality!
 
-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.