TheBloke/sqlcoder2-GGUF

1---2base_model: defog/sqlcoder23inference: false4language:5- en6license: other7model_creator: Defog.ai8model_name: Sqlcoder29model_type: starcoder10pipeline_tag: text-generation11prompt_template: "## Task\nGenerate a SQL query to answer the following question:\n\12  `{prompt}`\n\n### Database Schema\nThis query will run on a database whose schema\13  \ is represented in this string:\nCREATE TABLE products (\n  product_id INTEGER\14  \ PRIMARY KEY, -- Unique ID for each product\n  name VARCHAR(50), -- Name of the\15  \ product\n  price DECIMAL(10,2), -- Price of each unit of the product\n  quantity\16  \ INTEGER  -- Current quantity in stock\n);\n\nCREATE TABLE sales (\n  sale_id INTEGER\17  \ PRIMARY KEY, -- Unique ID for each sale\n  product_id INTEGER, -- ID of product\18  \ sold\n  customer_id INTEGER,  -- ID of customer who made purchase\n  salesperson_id\19  \ INTEGER, -- ID of salesperson who made the sale\n  sale_date DATE, -- Date the\20  \ sale occurred\n  quantity INTEGER -- Quantity of product sold\n);\n\n-- sales.product_id\21  \ can be joined with products.product_id\n\n### SQL\nGiven the database schema,\22  \ here is the SQL query that answers `{prompt}`:\n```sql\n"23quantized_by: TheBloke24tags:25- code26---27 28<!-- header start -->29<!-- 200823 -->30<div style="width: auto; margin-left: auto; margin-right: auto">31<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">32</div>33<div style="display: flex; justify-content: space-between; width: 100%;">34    <div style="display: flex; flex-direction: column; align-items: flex-start;">35        <p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>36    </div>37    <div style="display: flex; flex-direction: column; align-items: flex-end;">38        <p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>39    </div>40</div>41<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>42<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">43<!-- header end -->44 45# Sqlcoder2 - GGUF46- Model creator: [Defog.ai](https://huggingface.co/defog)47- Original model: [Sqlcoder2](https://huggingface.co/defog/sqlcoder2)48 49<!-- description start -->50## Description51 52This repo contains GGUF format model files for [Defog.ai's Sqlcoder2](https://huggingface.co/defog/sqlcoder2).53 54<!-- description end -->55<!-- README_GGUF.md-about-gguf start -->56### About GGUF57 58GGUF is a new format introduced by the llama.cpp team on August 21st 2023. It is a replacement for GGML, which is no longer supported by llama.cpp.59 60Here is an incomplate list of clients and libraries that are known to support GGUF:61 62* [llama.cpp](https://github.com/ggerganov/llama.cpp). The source project for GGUF. Offers a CLI and a server option.63* [text-generation-webui](https://github.com/oobabooga/text-generation-webui), the most widely used web UI, with many features and powerful extensions. Supports GPU acceleration.64* [KoboldCpp](https://github.com/LostRuins/koboldcpp), a fully featured web UI, with GPU accel across all platforms and GPU architectures. Especially good for story telling.65* [LM Studio](https://lmstudio.ai/), an easy-to-use and powerful local GUI for Windows and macOS (Silicon), with GPU acceleration.66* [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui), a great web UI with many interesting and unique features, including a full model library for easy model selection.67* [Faraday.dev](https://faraday.dev/), an attractive and easy to use character-based chat GUI for Windows and macOS (both Silicon and Intel), with GPU acceleration.68* [ctransformers](https://github.com/marella/ctransformers), a Python library with GPU accel, LangChain support, and OpenAI-compatible AI server.69* [llama-cpp-python](https://github.com/abetlen/llama-cpp-python), a Python library with GPU accel, LangChain support, and OpenAI-compatible API server.70* [candle](https://github.com/huggingface/candle), a Rust ML framework with a focus on performance, including GPU support, and ease of use.71 72<!-- README_GGUF.md-about-gguf end -->73<!-- repositories-available start -->74## Repositories available75 76* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/sqlcoder2-AWQ)77* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/sqlcoder2-GPTQ)78* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/sqlcoder2-GGUF)79* [Defog.ai's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/defog/sqlcoder2)80<!-- repositories-available end -->81 82<!-- prompt-template start -->83## Prompt template: Sqlcoder84 85```86## Task87Generate a SQL query to answer the following question:88`{prompt}`89 90### Database Schema91This query will run on a database whose schema is represented in this string:92CREATE TABLE products (93  product_id INTEGER PRIMARY KEY, -- Unique ID for each product94  name VARCHAR(50), -- Name of the product95  price DECIMAL(10,2), -- Price of each unit of the product96  quantity INTEGER  -- Current quantity in stock97);98 99CREATE TABLE sales (100  sale_id INTEGER PRIMARY KEY, -- Unique ID for each sale101  product_id INTEGER, -- ID of product sold102  customer_id INTEGER,  -- ID of customer who made purchase103  salesperson_id INTEGER, -- ID of salesperson who made the sale104  sale_date DATE, -- Date the sale occurred105  quantity INTEGER -- Quantity of product sold106);107 108-- sales.product_id can be joined with products.product_id109 110### SQL111Given the database schema, here is the SQL query that answers `{prompt}`:112```sql113 114```115 116<!-- prompt-template end -->117 118 119<!-- compatibility_gguf start -->120## Compatibility121 122These quantised GGUFv2 files are compatible with llama.cpp from August 27th onwards, as of commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221)123 124They are also compatible with many third party UIs and libraries - please see the list at the top of this README.125 126## Explanation of quantisation methods127<details>128  <summary>Click to see details</summary>129 130The new methods available are:131* GGML_TYPE_Q2_K - "type-1" 2-bit quantization in super-blocks containing 16 blocks, each block having 16 weight. Block scales and mins are quantized with 4 bits. This ends up effectively using 2.5625 bits per weight (bpw)132* GGML_TYPE_Q3_K - "type-0" 3-bit quantization in super-blocks containing 16 blocks, each block having 16 weights. Scales are quantized with 6 bits. This end up using 3.4375 bpw.133* GGML_TYPE_Q4_K - "type-1" 4-bit quantization in super-blocks containing 8 blocks, each block having 32 weights. Scales and mins are quantized with 6 bits. This ends up using 4.5 bpw.134* GGML_TYPE_Q5_K - "type-1" 5-bit quantization. Same super-block structure as GGML_TYPE_Q4_K resulting in 5.5 bpw135* GGML_TYPE_Q6_K - "type-0" 6-bit quantization. Super-blocks with 16 blocks, each block having 16 weights. Scales are quantized with 8 bits. This ends up using 6.5625 bpw136 137Refer to the Provided Files table below to see what files use which methods, and how.138</details>139<!-- compatibility_gguf end -->140 141<!-- README_GGUF.md-provided-files start -->142## Provided files143 144| Name | Quant method | Bits | Size | Max RAM required | Use case |145| ---- | ---- | ---- | ---- | ---- | ----- |146| [sqlcoder2.Q2_K.gguf](https://huggingface.co/TheBloke/sqlcoder2-GGUF/blob/main/sqlcoder2.Q2_K.gguf) | Q2_K | 2 | 6.73 GB| 9.23 GB | smallest, significant quality loss - not recommended for most purposes |147| [sqlcoder2.Q3_K_S.gguf](https://huggingface.co/TheBloke/sqlcoder2-GGUF/blob/main/sqlcoder2.Q3_K_S.gguf) | Q3_K_S | 3 | 6.93 GB| 9.43 GB | very small, high quality loss |148| [sqlcoder2.Q3_K_M.gguf](https://huggingface.co/TheBloke/sqlcoder2-GGUF/blob/main/sqlcoder2.Q3_K_M.gguf) | Q3_K_M | 3 | 8.18 GB| 10.68 GB | very small, high quality loss |149| [sqlcoder2.Q4_0.gguf](https://huggingface.co/TheBloke/sqlcoder2-GGUF/blob/main/sqlcoder2.Q4_0.gguf) | Q4_0 | 4 | 8.99 GB| 11.49 GB | legacy; small, very high quality loss - prefer using Q3_K_M |150| [sqlcoder2.Q4_K_S.gguf](https://huggingface.co/TheBloke/sqlcoder2-GGUF/blob/main/sqlcoder2.Q4_K_S.gguf) | Q4_K_S | 4 | 9.06 GB| 11.56 GB | small, greater quality loss |151| [sqlcoder2.Q3_K_L.gguf](https://huggingface.co/TheBloke/sqlcoder2-GGUF/blob/main/sqlcoder2.Q3_K_L.gguf) | Q3_K_L | 3 | 9.08 GB| 11.58 GB | small, substantial quality loss |152| [sqlcoder2.Q4_K_M.gguf](https://huggingface.co/TheBloke/sqlcoder2-GGUF/blob/main/sqlcoder2.Q4_K_M.gguf) | Q4_K_M | 4 | 9.96 GB| 12.46 GB | medium, balanced quality - recommended |153| [sqlcoder2.Q5_0.gguf](https://huggingface.co/TheBloke/sqlcoder2-GGUF/blob/main/sqlcoder2.Q5_0.gguf) | Q5_0 | 5 | 10.93 GB| 13.43 GB | legacy; medium, balanced quality - prefer using Q4_K_M |154| [sqlcoder2.Q5_K_S.gguf](https://huggingface.co/TheBloke/sqlcoder2-GGUF/blob/main/sqlcoder2.Q5_K_S.gguf) | Q5_K_S | 5 | 10.93 GB| 13.43 GB | large, low quality loss - recommended |155| [sqlcoder2.Q5_K_M.gguf](https://huggingface.co/TheBloke/sqlcoder2-GGUF/blob/main/sqlcoder2.Q5_K_M.gguf) | Q5_K_M | 5 | 11.54 GB| 14.04 GB | large, very low quality loss - recommended |156| [sqlcoder2.Q6_K.gguf](https://huggingface.co/TheBloke/sqlcoder2-GGUF/blob/main/sqlcoder2.Q6_K.gguf) | Q6_K | 6 | 12.99 GB| 15.49 GB | very large, extremely low quality loss |157| [sqlcoder2.Q8_0.gguf](https://huggingface.co/TheBloke/sqlcoder2-GGUF/blob/main/sqlcoder2.Q8_0.gguf) | Q8_0 | 8 | 16.82 GB| 19.32 GB | very large, extremely low quality loss - not recommended |158 159**Note**: the above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead.160 161 162 163<!-- README_GGUF.md-provided-files end -->164 165<!-- README_GGUF.md-how-to-download start -->166## How to download GGUF files167 168**Note for manual downloaders:** You almost never want to clone the entire repo! Multiple different quantisation formats are provided, and most users only want to pick and download a single file.169 170The following clients/libraries will automatically download models for you, providing a list of available models to choose from:171- LM Studio172- LoLLMS Web UI173- Faraday.dev174 175### In `text-generation-webui`176 177Under Download Model, you can enter the model repo: TheBloke/sqlcoder2-GGUF and below it, a specific filename to download, such as: sqlcoder2.Q4_K_M.gguf.178 179Then click Download.180 181### On the command line, including multiple files at once182 183I recommend using the `huggingface-hub` Python library:184 185```shell186pip3 install huggingface-hub187```188 189Then you can download any individual model file to the current directory, at high speed, with a command like this:190 191```shell192huggingface-cli download TheBloke/sqlcoder2-GGUF sqlcoder2.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False193```194 195<details>196  <summary>More advanced huggingface-cli download usage</summary>197 198You can also download multiple files at once with a pattern:199 200```shell201huggingface-cli download TheBloke/sqlcoder2-GGUF --local-dir . --local-dir-use-symlinks False --include='*Q4_K*gguf'202```203 204For more documentation on downloading with `huggingface-cli`, please see: [HF -> Hub Python Library -> Download files -> Download from the CLI](https://huggingface.co/docs/huggingface_hub/guides/download#download-from-the-cli).205 206To accelerate downloads on fast connections (1Gbit/s or higher), install `hf_transfer`:207 208```shell209pip3 install hf_transfer210```211 212And set environment variable `HF_HUB_ENABLE_HF_TRANSFER` to `1`:213 214```shell215HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/sqlcoder2-GGUF sqlcoder2.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False216```217 218Windows Command Line users: You can set the environment variable by running `set HF_HUB_ENABLE_HF_TRANSFER=1` before the download command.219</details>220<!-- README_GGUF.md-how-to-download end -->221 222<!-- README_GGUF.md-how-to-run start -->223## Example `llama.cpp` command224 225Make sure you are using `llama.cpp` from commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221) or later.226 227```shell228./main -ngl 32 -m sqlcoder2.Q4_K_M.gguf --color -c 2048 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "## Task\nGenerate a SQL query to answer the following question:\n`{prompt}`\n\n### Database Schema\nThis query will run on a database whose schema is represented in this string:\nCREATE TABLE products (\n  product_id INTEGER PRIMARY KEY, -- Unique ID for each product\n  name VARCHAR(50), -- Name of the product\n  price DECIMAL(10,2), -- Price of each unit of the product\n  quantity INTEGER  -- Current quantity in stock\n);\n\nCREATE TABLE sales (\n  sale_id INTEGER PRIMARY KEY, -- Unique ID for each sale\n  product_id INTEGER, -- ID of product sold\n  customer_id INTEGER,  -- ID of customer who made purchase\n  salesperson_id INTEGER, -- ID of salesperson who made the sale\n  sale_date DATE, -- Date the sale occurred\n  quantity INTEGER -- Quantity of product sold\n);\n\n-- sales.product_id can be joined with products.product_id\n\n### SQL\nGiven the database schema, here is the SQL query that answers `{prompt}`:\n```sql"229```230 231Change `-ngl 32` to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.232 233Change `-c 2048` to the desired sequence length. For extended sequence models - eg 8K, 16K, 32K - the necessary RoPE scaling parameters are read from the GGUF file and set by llama.cpp automatically.234 235If you want to have a chat-style conversation, replace the `-p <PROMPT>` argument with `-i -ins`236 237For other parameters and how to use them, please refer to [the llama.cpp documentation](https://github.com/ggerganov/llama.cpp/blob/master/examples/main/README.md)238 239## How to run in `text-generation-webui`240 241Further instructions here: [text-generation-webui/docs/llama.cpp.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/llama.cpp.md).242 243## How to run from Python code244 245You can use GGUF models from Python using the [llama-cpp-python](https://github.com/abetlen/llama-cpp-python) or [ctransformers](https://github.com/marella/ctransformers) libraries.246 247### How to load this model in Python code, using ctransformers248 249#### First install the package250 251Run one of the following commands, according to your system:252 253```shell254# Base ctransformers with no GPU acceleration255pip install ctransformers256# Or with CUDA GPU acceleration257pip install ctransformers[cuda]258# Or with AMD ROCm GPU acceleration (Linux only)259CT_HIPBLAS=1 pip install ctransformers --no-binary ctransformers260# Or with Metal GPU acceleration for macOS systems only261CT_METAL=1 pip install ctransformers --no-binary ctransformers262```263 264#### Simple ctransformers example code265 266```python267from ctransformers import AutoModelForCausalLM268 269# Set gpu_layers to the number of layers to offload to GPU. Set to 0 if no GPU acceleration is available on your system.270llm = AutoModelForCausalLM.from_pretrained("TheBloke/sqlcoder2-GGUF", model_file="sqlcoder2.Q4_K_M.gguf", model_type="starcoder", gpu_layers=50)271 272print(llm("AI is going to"))273```274 275## How to use with LangChain276 277Here are guides on using llama-cpp-python and ctransformers with LangChain:278 279* [LangChain + llama-cpp-python](https://python.langchain.com/docs/integrations/llms/llamacpp)280* [LangChain + ctransformers](https://python.langchain.com/docs/integrations/providers/ctransformers)281 282<!-- README_GGUF.md-how-to-run end -->283 284<!-- footer start -->285<!-- 200823 -->286## Discord287 288For further support, and discussions on these models and AI in general, join us at:289 290[TheBloke AI's Discord server](https://discord.gg/theblokeai)291 292## Thanks, and how to contribute293 294Thanks to the [chirper.ai](https://chirper.ai) team!295 296Thanks to Clay from [gpus.llm-utils.org](llm-utils)!297 298I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.299 300If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.301 302Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.303 304* Patreon: https://patreon.com/TheBlokeAI305* Ko-Fi: https://ko-fi.com/TheBlokeAI306 307**Special thanks to**: Aemon Algiz.308 309**Patreon special mentions**: Pierre Kircher, Stanislav Ovsiannikov, Michael Levine, Eugene Pentland, Andrey, 준교 김, Randy H, Fred von Graf, Artur Olbinski, Caitlyn Gatomon, terasurfer, Jeff Scroggin, James Bentley, Vadim, Gabriel Puliatti, Harry Royden McLaughlin, Sean Connelly, Dan Guido, Edmond Seymore, Alicia Loh, subjectnull, AzureBlack, Manuel Alberto Morcote, Thomas Belote, Lone Striker, Chris Smitley, Vitor Caleffi, Johann-Peter Hartmann, Clay Pascal, biorpg, Brandon Frisco, sidney chen, transmissions 11, Pedro Madruga, jinyuan sun, Ajan Kanaga, Emad Mostaque, Trenton Dambrowitz, Jonathan Leane, Iucharbius, usrbinkat, vamX, George Stoitzev, Luke Pendergrass, theTransient, Olakabola, Swaroop Kallakuri, Cap'n Zoog, Brandon Phillips, Michael Dempsey, Nikolai Manek, danny, Matthew Berman, Gabriel Tamborski, alfie_i, Raymond Fosdick, Tom X Nguyen, Raven Klaugh, LangChain4j, Magnesian, Illia Dulskyi, David Ziegler, Mano Prime, Luis Javier Navarrete Lozano, Erik Bjäreholt, 阿明, Nathan Dryer, Alex, Rainer Wilmers, zynix, TL, Joseph William Delisle, John Villwock, Nathan LeClaire, Willem Michiel, Joguhyik, GodLy, OG, Alps Aficionado, Jeffrey Morgan, ReadyPlayerEmma, Tiffany J. Kim, Sebastain Graf, Spencer Kim, Michael Davis, webtim, Talal Aujan, knownsqashed, John Detwiler, Imad Khwaja, Deo Leter, Jerry Meng, Elijah Stavena, Rooh Singh, Pieter, SuperWojo, Alexandros Triantafyllidis, Stephen Murray, Ai Maven, ya boyyy, Enrico Ros, Ken Nordquist, Deep Realms, Nicholas, Spiking Neurons AB, Elle, Will Dee, Jack West, RoA, Luke @flexchar, Viktor Bowallius, Derek Yates, Subspace Studios, jjj, Toran Billups, Asp the Wyvern, Fen Risland, Ilya, NimbleBox.ai, Chadd, Nitin Borwankar, Emre, Mandus, Leonard Tan, Kalila, K, Trailburnt, S_X, Cory Kujawski310 311 312Thank you to all my generous patrons and donaters!313 314And thank you again to a16z for their generous grant.315 316<!-- footer end -->317 318<!-- original-model-card start -->319# Original model card: Defog.ai's Sqlcoder2320 321# Defog SQLCoder322Defog's SQLCoder is a state-of-the-art LLM for converting natural language questions to SQL queries.323 324[Interactive Demo](https://defog.ai/sqlcoder-demo/) | [🤗 HF Repo](https://huggingface.co/defog/sqlcoder2) | [♾️ Colab](https://colab.research.google.com/drive/1z4rmOEiFkxkMiecAWeTUlPl0OmKgfEu7?usp=sharing) | [🐦 Twitter](https://twitter.com/defogdata)325 326## TL;DR327SQLCoder is a 15B parameter model that outperforms `gpt-3.5-turbo` for natural language to SQL generation tasks on our [sql-eval](https://github.com/defog-ai/sql-eval) framework, and significantly outperforms all popular open-source models. When fine-tuned on a given schema, it also outperforms `gpt-4`328 329SQLCoder is fine-tuned on a base StarCoder model.330 331## Results on novel datasets not seen in training332| model   | perc_correct |333|-|-|334| gpt4-2023-10-04    | 82.0 |335| defog-sqlcoder2    | 74.5 |336| gpt4-2023-08-28    | 74.0 |337| defog-sqlcoder-7b  | 71.0 |338| gpt-3.5-2023-10-04 | 66.0 |339| claude-2           | 64.5 |340| gpt-3.5-2023-08-28 | 61.0 |341| claude_instant_1   | 61.0 |342| text-davinci-003   | 52.5 |343 344## License345The code in this repo (what little there is of it) is Apache-2 licensed. The model weights have a `CC BY-SA 4.0` license, with additional responsible use restrictions added. The TL;DR is that you can use and modify the model for any purpose – including commercial use. However, if you modify the weights (for example, by fine-tuning), you must open-source your modified weights under the same license terms.346 347## Training348Defog was trained on more than 20,000 human-curated questions. These questions were based on 10 different schemas. None of the schemas in the training data were included in our evaluation framework.349 350You can read more about our [training approach](https://defog.ai/blog/open-sourcing-sqlcoder2-7b/) and [evaluation framework](https://defog.ai/blog/open-sourcing-sqleval/).351 352## Results by question category353We classified each generated question into one of 5 categories. The table displays the percentage of questions answered correctly by each model, broken down by category.354| query_category   |   gpt-4 |   sqlcoder2-15b |   sqlcoder-7b |   gpt-3.5 |   claude-2 |   claude-instant |   gpt-3 |355|:-----------------|--------:|----------------:|--------------:|----------:|-----------:|-----------------:|--------:|356| date             |    72   |            76   |          64   |      68   |       52   |             48   |    32   |357| group_by         |    91.4 |            80   |          82.9 |      77.1 |       71.4 |             71.4 |    71.4 |358| order_by         |    82.9 |            77.1 |          74.3 |      68.6 |       74.3 |             74.3 |    68.6 |359| ratio            |    80   |            60   |          54.3 |      37.1 |       57.1 |             45.7 |    25.7 |360| join             |    82.9 |            77.1 |          74.3 |      71.4 |       65.7 |             62.9 |    57.1 |361| where            |    80   |            77.1 |          74.3 |      74.3 |       62.9 |             60   |    54.3 |362 363## Using SQLCoder364You can use SQLCoder via the `transformers` library by downloading our model weights from the Hugging Face repo. We have added sample code for [inference](./inference.py) on a [sample database schema](./metadata.sql).365```bash366python inference.py -q "Question about the sample database goes here"367 368# Sample question:369# Do we get more revenue from customers in New York compared to customers in San Francisco? Give me the total revenue for each city, and the difference between the two.370```371 372You can also use a demo on our website [here](https://defog.ai/sqlcoder-demo), or run SQLCoder in Colab [here](https://colab.research.google.com/drive/13BIKsqHnPOBcQ-ba2p77L5saiepTIwu0#scrollTo=ZpbVgVHMkJvC)373 374## Hardware Requirements375SQLCoder has been tested on an A100 40GB GPU with `bfloat16` weights. You can also load an 8-bit and 4-bit quantized version of the model on consumer GPUs with 20GB or more of memory – like RTX 4090, RTX 3090, and Apple M2 Pro, M2 Max, or M2 Ultra Chips with 20GB or more of memory.376 377## Todo378 379- [x] Open-source the v1 model weights380- [x] Train the model on more data, with higher data variance381- [ ] Tune the model further with Reward Modelling and RLHF382- [ ] Pretrain a model from scratch that specializes in SQL analysis383 384<!-- original-model-card end -->385