mlc-ai
/

Mistral-7B-Instruct-v0.2-q3f16_1-MLC

Model card Files Files and versions

davidpissarra commited on Jan 12, 2024

Commit

c925345

·

1 Parent(s): 1146aeb

change prefill_chunk_size

Files changed (1) hide show

mlc-chat-config.json +2 -2

mlc-chat-config.json CHANGED Viewed

@@ -13,14 +13,14 @@
     "num_key_value_heads": 8,
     "head_dim": 128,
     "sliding_window_size": 1024,
-    "prefill_chunk_size": 1024,
     "attention_sink_size": 4,
     "tensor_parallel_shards": 1
   },
   "vocab_size": 32000,
   "context_window_size": -1,
   "sliding_window_size": 1024,
-  "prefill_chunk_size": 1024,
   "attention_sink_size": 4,
   "tensor_parallel_shards": 1,
   "mean_gen_len": 128,

     "num_key_value_heads": 8,
     "head_dim": 128,
     "sliding_window_size": 1024,
+    "prefill_chunk_size": 128,
     "attention_sink_size": 4,
     "tensor_parallel_shards": 1
   },
   "vocab_size": 32000,
   "context_window_size": -1,
   "sliding_window_size": 1024,
+  "prefill_chunk_size": 128,
   "attention_sink_size": 4,
   "tensor_parallel_shards": 1,
   "mean_gen_len": 128,