From 1a99b95c4332d439fa50fdf78bd35e58fff6ec8c Mon Sep 17 00:00:00 2001
From: Brian Lee <brian@lilacml.com>
Date: Wed, 24 Jan 2024 11:06:14 -0500
Subject: [PATCH] Add mosaic-instruct-v3 (#1116)

---
 lilac_hf_space.yml | 22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)

diff --git a/lilac_hf_space.yml b/lilac_hf_space.yml
index dc4fabc5..8a1e50a7 100644
--- a/lilac_hf_space.yml
+++ b/lilac_hf_space.yml
@@ -122,6 +122,23 @@ datasets:
         path:
           - question
 
+  - namespace: lilac
+    name: mosaic-instruct-v3
+    source:
+      dataset_name: mosaicml/instruct-v3
+      source_name: huggingface
+    settings:
+      tags: [datasets]
+      ui:
+        media_paths:
+          - prompt
+          - response
+        markdown_paths: []
+    embeddings:
+      - embedding: gte-small
+        path:
+          - prompt
+
   ## Eval datasets
   - namespace: lilac
     name: MMLU
@@ -387,6 +404,11 @@ clusters:
     input_path:
       - original-instruction
 
+  - dataset_namespace: lilac
+    dataset_name: mosaic-instruct-v3
+    input_path:
+      - prompt
+
   ## Eval datasets
   - dataset_namespace: lilac
     dataset_name: MMLU