How to use from
Unsloth Studio
Install Unsloth Studio (macOS, Linux, WSL)
curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for jastorj/couchmind-rlt-v5.8.1_rl-cw-8K-16bit to start chatting
Install Unsloth Studio (Windows)
irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for jastorj/couchmind-rlt-v5.8.1_rl-cw-8K-16bit to start chatting
Using HuggingFace Spaces for Unsloth
# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for jastorj/couchmind-rlt-v5.8.1_rl-cw-8K-16bit to start chatting
Load model with FastModel
pip install unsloth
from unsloth import FastModel
model, tokenizer = FastModel.from_pretrained(
    model_name="jastorj/couchmind-rlt-v5.8.1_rl-cw-8K-16bit",
    max_seq_length=2048,
)
Quick Links

jastorj/couchmind-v5.8.1_cold_start-cw-30K-16bit Fine-tuned for NL2SQL++ v5.8.1_rl (RLT)

This model is a reinforcement-learning fine-tuned version of jastorj/couchmind-v5.8.1_cold_start-cw-30K-16bit on the NL2SQL++ v5.8.1_rl dataset.

Model Details

  • Base Model: jastorj/couchmind-v5.8.1_cold_start-cw-30K-16bit
  • Task: Text-to-SQL generation
  • Dataset: NL2SQL++ v5.8.1_rl
  • Fine-tuning Method: GRPO (Group Relative Policy Optimisation) with Unsloth
  • Quantization: 16-bit merged weights
  • Training Dataset Size: 48 examples
  • Validation Dataset Size: 0 examples

Full training configuration: see training_args.json in this repo.

Train Dataset Example

<|im_start|>system
You are a helpful AI Assistant that provides well-reasoned and detailed responses. You first
think about the reasoning process as an internal monologue and then provide the user with
the answer. Respond in the following format: <think>
...
</think>
<answer>
...
</answer><|im_end|>
<|im_start|>user
You are a Couchbase SQL++ query expert. Given a database schema and a natural language question, generate a syntactically valid SQL++ query that precisely answers the question.

Rules:
- SELECT only the columns explicitly asked for — nothing more, nothing less
- Use the exact bucket, scope, and collection names provided in the database schema

Bucket Name:
`spider_training_bucket`

Scope Name:
`chinook_1`

Database Schema:
{"`spider_training_bucket`.`chinook_1`.`MediaType`": {"properties": {"name": {"samples": ["MPEG audio file", "AAC audio file"], "type": "string"}, "mediatypeid": {"samples": [1, 2], "type": "number"}}, "type": "object"}, "`spider_training_bucket`.`chinook_1`.`PlaylistTrack`": {"properties": {"trackid": {"samples": [394, 1138], "type": "number"}, "playlistid": {"samples": [3, 1], "type": "number"}}, "type": "object"}, "`spider_training_bucket`.`chinook_1`.`Employee`": {"properties": {"firstname": {"samples": ["Andrew", "Jane"], "type": "string"}, "birthdate": {"samples": ["1958-12-08 00:00:00", "1947-09-19 00:00:00"], "type": "string"}, "state": {"samples": ["AB"], "type": "string"}, "country": {"samples": ["Canada"], "type": "string"}, "employeeid": {"samples": [1, 2], "type": "number"}, "postalcode": {"samples": ["T1K 5N8", "T1H 1Y8"], "type": "string"}, "address": {"samples": ["1111 6 Ave SW", "11120 Jasper Ave NW"], "type": "string"}, "fax": {"samples": ["+1 (403) 262-3322", "+1 (403) 246-9899"], "type": "string"}, "hiredate": {"samples": ["2002-05-01 00:00:00", "2002-04-01 00:00:00"], "type": "string"}, "phone": {"samples": ["+1 (403) 246-9887", "+1 (403) 262-3443"], "type": "string"}, "title": {"samples": ["General Manager", "IT Manager"], "type": "string"}, "city": {"samples": ["Edmonton", "Calgary"], "type": "string"}, "lastname": {"samples": ["Adams", "Callahan"], "type": "string"}, "reportsto": {"samples": [2, 1], "type": ["null", "number"]}, "email": {"samples": ["[email protected]", "[email protected]"], "type": "string"}}, "type": "object"}, "`spider_training_bucket`.`chinook_1`.`Genre`": {"properties": {"name": {"samples": ["Bossa Nova", "Blues"], "type": "string"}, "genreid": {"samples": [5, 1], "type": "number"}}, "type": "object"}, "`spider_training_bucket`.`chinook_1`.`InvoiceLine`": {"properties": {"invoicelineid": {"samples": [497, 230], "type": "number"}, "unitprice": {"samples": [1.99, 0.99], "type": "number"}, "trackid": {"samples": [1396, 450], "type": "number"}, "quantity": {"samples": [1], "type": "number"}, "invoiceid": {"samples": [43, 93], "type": "number"}}, "type": "object"}, "`spider_training_bucket`.`chinook_1`.`Artist`": {"properties": {"name": {"samples": ["Chris Cornell", "Antal Dor\u00e1ti & London Symphon..."], "type": "string"}, "artistid": {"samples": [75, 76], "type": "number"}}, "type": "object"}, "`spider_training_bucket`.`chinook_1`.`Album`": {"properties": {"albumid": {"samples": [51, 84], "type": "number"}, "title": {"samples": ["20th Century Masters - The Mil...", "By The Way"], "type": "string"}, "artistid": {"samples": [50, 69], "type": "number"}}, "type": "object"}, "`spider_training_bucket`.`chinook_1`.`Invoice`": {"properties": {"customerid": {"samples": [5, 6], "type": "number"}, "invoiceid": {"samples": [46, 30], "type": "number"}, "billingaddress": {"samples": ["110 Raeburn Pl", "12,Community Centre"], "type": "string"}, "total": {"samples": [1.98, 0.99], "type": "number"}, "billingstate": {"samples": ["DF", "BC"], "type": ["null", "string"]}, "billingpostalcode": {"samples": ["110017", "10779"], "type": ["null", "string"]}, "billingcity": {"samples": ["Bangalore", "Berlin"], "type": "string"}, "invoicedate": {"samples": ["2009-07-11 00:00:00", "2009-05-06 00:00:00"], "type": "string"}, "billingcountry": {"samples": ["Argentina", "Austria"], "type": "string"}}, "type": "object"}, "`spider_training_bucket`.`chinook_1`.`Playlist`": {"properties": {"name": {"samples": ["Audiobooks", "90\u2019s Music"], "type": "string"}, "playlistid": {"samples": [2, 4], "type": "number"}}, "type": "object"}, "`spider_training_bucket`.`chinook_1`.`Customer`": {"properties": {"firstname": {"samples": ["Daan", "Bj\u00f8rn"], "type": "string"}, "email": {"samples": ["[email protected]", "[email protected]"], "type": "string"}, "phone": {"samples": ["+33 01 47 42 71 71", "+32 02 219 03 03"], "type": ["null", "string"]}, "lastname": {"samples": ["Hansen", "Guti\u00e9rrez"], "type": "string"}, "customerid": {"samples": [2, 4], "type": "number"}, "country": {"samples": ["Argentina", "Belgium"], "type": "string"}, "city": {"samples": ["Berlin", "Bangalore"], "type": "string"}, "supportrepid": {"samples": [3, 4], "type": "number"}, "company": {"samples": ["Apple Inc.", "Banco do Brasil S.A."], "type": ["null", "string"]}, "address": {"samples": ["3,Raj Bhavan Road", "110 Raeburn Pl"], "type": "string"}, "fax": {"samples": ["+1 (425) 882-8081", "+1 (408) 996-1011"], "type": ["null", "string"]}, "postalcode": {"samples": ["1000", "0171"], "type": ["null", "string"]}, "state": {"samples": ["IL", "CA"], "type": ["null", "string"]}}, "type": "object"}, "`spider_training_bucket`.`chinook_1`.`Track`": {"properties": {"albumid": {"samples": [7, 26], "type": "number"}, "genreid": {"samples": [1, 3], "type": "number"}, "trackid": {"samples": [55, 295], "type": "number"}, "mediatypeid": {"samples": [2, 1], "type": "number"}, "unitprice": {"samples": [1.99, 0.99], "type": "number"}, "bytes": {"samples": [5566464, 2775071], "type": "number"}, "name": {"samples": ["Bad", "Concilia\u00e7\u00e3o"], "type": "string"}, "composer": {"samples": ["Bruce Dickinson/Janick Gers", "Alcides Dias Lopes"], "type": ["null", "string"]}, "milliseconds": {"samples": [172710, 180035], "type": "number"}}, "type": "object"}}

This schema describes the structure of the data in the specified bucket and scope. It includes information about the collections, fields, and their data types.

Question:
Please list all album titles in alphabetical order.<|im_end|>
<|im_start|>assistant
Downloads last month
29
Safetensors
Model size
8B params
Tensor type
BF16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for jastorj/couchmind-rlt-v5.8.1_rl-cw-8K-16bit