p2o6e100
/

ASVDLlama3.1-4B-Instruct-v1-q0f16-MLC

Model card Files Files and versions

p2o6e100 commited on Nov 15, 2024

Commit

a0a33e5

·

1 Parent(s): 362fb4b

modify config

Files changed (2) hide show

.DS_Store +0 -0
mlc-chat-config.json +4 -4

.DS_Store ADDED Viewed

Binary file (14.3 kB). View file

mlc-chat-config.json CHANGED Viewed

@@ -18,8 +18,8 @@
       "original_max_position_embeddings": 8192,
       "rope_type": "llama3"
     },
-    "context_window_size": 1024,
-    "prefill_chunk_size": 1024,
     "num_key_value_heads": 7,
     "head_dim": 128,
     "tensor_parallel_shards": 1,
@@ -164,9 +164,9 @@
     "max_batch_size": 1
   },
   "vocab_size": 77209,
-  "context_window_size": 1024,
   "sliding_window_size": -1,
-  "prefill_chunk_size": 1024,
   "attention_sink_size": -1,
   "tensor_parallel_shards": 1,
   "pipeline_parallel_stages": 1,

       "original_max_position_embeddings": 8192,
       "rope_type": "llama3"
     },
+    "context_window_size": 768,
+    "prefill_chunk_size": 768,
     "num_key_value_heads": 7,
     "head_dim": 128,
     "tensor_parallel_shards": 1,
     "max_batch_size": 1
   },
   "vocab_size": 77209,
+  "context_window_size": 768,
   "sliding_window_size": -1,
+  "prefill_chunk_size": 768,
   "attention_sink_size": -1,
   "tensor_parallel_shards": 1,
   "pipeline_parallel_stages": 1,