textsoketlabs/bhasha-sftsftinstruction-tuningmultilingualhindibengaligujaratiindicllm-trainingfine-tuningparquet
Bhasha SFT — 13M+ Multilingual Instruction-Response Pairs (Hindi, Bengali, Gujarati, English)
About this data
13M+ instruction-response pairs across Hindi, Bengali, Gujarati, and English for supervised fine-tuning of multilingual LLMs. Mix of human-annotated and synthetic data from open-source SFT collections, curated by Soket AI Labs.
Schema
| Name | Type | Description |
|---|---|---|
| doc_id | VARCHAR | Unique identifier for the instruction-response document |
| num_turns | BIGINT | Count of conversation turns (user-assistant exchanges) in the messages |
| messages | STRUCT("content" VARCHAR, "role" VARCHAR)[] | Array of alternating user and assistant messages with content and role fields |
| language | VARCHAR | ISO 639-3 language code (ben, hin, guj, eng) |
| script | VARCHAR | Writing system code (Beng, Deva, Gujr, Latn) |
| source | VARCHAR | Name of upstream dataset this instance originated from |
| task | VARCHAR | Task category such as question-answering, summarization, translation, or classification |
| topic | VARCHAR | Subject domain or topic area covered by the instruction-response pair |
Sample Data
Preview a sample of the data before downloading.
Free
Open dataset
Quality: No ratings
0 downloads
Seller: DataBazaar
Agent? No sign-up needed →
For AI Agents
Via MCP Server
# 1. Add to your agent's MCP config (claude_desktop_config.json or similar):
{
"mcpServers": {
"databazaar": { "command": "npx", "args": ["databazaar-mcp"] }
}
}
# 2. Your agent can then call:
search_datasets({ query: "Bhasha SFT — 13M+ Multilingual" })
// Found: 587f81d5-4838-4d5e-9b6a-9dd651def675
get_download_url({ dataset_id: "587f81d5-4838-4d5e-9b6a-9dd651def675" }) // free — no API key neededVia REST API
# Free dataset — no API key required: curl https://api.databazaar.io/datasets/587f81d5-4838-4d5e-9b6a-9dd651def675/download-url