Push to HF space

README.md

@@ -26,6 +26,7 @@ datasets:
   - lilacai/lilac-GSM8K-socratic
   - lilacai/lilac-WinoGrande
   - lilacai/lilac-databricks-dolly-15k-curated-en
+  - lilacai/lilac-mosaic-instruct-v3
   - lilacai/lilac-dolphin
 
 ---

data/lilac.yml

@@ -328,6 +328,21 @@ datasets:
             - '*'
       tags:
         - datasets
+  - namespace: lilac
+    name: mosaic-instruct-v3
+    source:
+      dataset_name: mosaicml/instruct-v3
+      source_name: huggingface
+    embeddings:
+      - path: prompt
+        embedding: gte-small
+    settings:
+      ui:
+        media_paths:
+          - prompt
+          - response
+      tags:
+        - datasets
   - namespace: lilac
     name: dolphin
     source:
@@ -377,7 +392,7 @@ clusters:
   - dataset_namespace: lilac
     dataset_name: lmsys-chat-1m
     input_selector:
-      format: openai_json
+      format: openai_conversation_json
       selector: user
     output_path: !!python/tuple
       - conversation__clusters
@@ -396,10 +411,21 @@ clusters:
     dataset_name: databricks-dolly-15k-curated-en
     input_path: !!python/tuple
       - original-instruction
+  - dataset_namespace: lilac
+    dataset_name: mosaic-instruct-v3
+    input_path: !!python/tuple
+      - prompt
   - dataset_namespace: lilac
     dataset_name: dolphin
     input_path: !!python/tuple
       - input
+  - dataset_namespace: lilac
+    dataset_name: UltraChat-200k
+    input_selector:
+      format: openai_json
+      selector: user
+    output_path: !!python/tuple
+      - messages__clusters
   - dataset_namespace: lilac
     dataset_name: MMLU
     input_path: !!python/tuple

dist/lilac-1337.0.0-py3-none-any.whl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3a4c40756512bfb589e2712849a5231a1d333de31ca87f55e096a98fe3d18275
-size 2481715
+oid sha256:52ff004683355decb00650d9ef844b561ad28cef12d278e4fe1f3470578a2b6e
+size 2481988