Tai Truong commited on
Commit
a8477f0
·
1 Parent(s): 9895f82

Add nb-llama-3.2-1B instruct model weights

Browse files
Files changed (1) hide show
  1. mlc-chat-config.json +4 -4
mlc-chat-config.json CHANGED
@@ -18,8 +18,8 @@
18
  "original_max_position_embeddings": 8192,
19
  "rope_type": "llama3"
20
  },
21
- "context_window_size": 131072,
22
- "prefill_chunk_size": 8192,
23
  "num_key_value_heads": 8,
24
  "head_dim": 64,
25
  "tensor_parallel_shards": 1,
@@ -27,9 +27,9 @@
27
  "max_batch_size": 128
28
  },
29
  "vocab_size": 128256,
30
- "context_window_size": 131072,
31
  "sliding_window_size": -1,
32
- "prefill_chunk_size": 8192,
33
  "attention_sink_size": -1,
34
  "tensor_parallel_shards": 1,
35
  "pipeline_parallel_stages": 1,
 
18
  "original_max_position_embeddings": 8192,
19
  "rope_type": "llama3"
20
  },
21
+ "context_window_size": 2048,
22
+ "prefill_chunk_size": 64,
23
  "num_key_value_heads": 8,
24
  "head_dim": 64,
25
  "tensor_parallel_shards": 1,
 
27
  "max_batch_size": 128
28
  },
29
  "vocab_size": 128256,
30
+ "context_window_size": 2048,
31
  "sliding_window_size": -1,
32
+ "prefill_chunk_size": 64,
33
  "attention_sink_size": -1,
34
  "tensor_parallel_shards": 1,
35
  "pipeline_parallel_stages": 1,