update debug

stanford-crfm · Jan 29, 2024 · d7238df · d7238df
1 parent 0faf92e
commit d7238df
Showing 1 changed file with 13 additions and 4 deletions.
diff --git a/config/data/rpv1_debug.yaml b/config/data/rpv1_debug.yaml
@@ -1,7 +1,16 @@
 cache_dir: gs://levanter-data/tokenized/redpajama_v1_llama_mixture
 rows_per_chunk: 4096
 tokenizer: "meta-llama/Llama-2-7b-hf"
-train_urls:
-  - gs://levanter-data/dev/redpajama/wikipedia/wiki.jsonl
-validation_urls:
-  - https://data.together.xyz/redpajama-data-1T/v1.0.0/c4/c4-train.01023-of-01024.jsonl
+configs:
+  # StackExchange:
+  #   train_urls:
+  #     - gs://levanter-data/dev/redpajama/stackexchange/stackexchange.jsonl
+  wikipedia:
+    train_urls:
+      - gs://levanter-data/dev/redpajama/wikipedia/wiki.jsonl
+    validation_urls:
+      - https://data.together.xyz/redpajama-data-1T/v1.0.0/c4/c4-train.01023-of-01024.jsonl
+train_weights:
+  # StackExchange: 
+  wikipedia: 24
+stop_strategy: all_exhausted