Update mlc-chat-config.json
Browse files- mlc-chat-config.json +69 -40
mlc-chat-config.json
CHANGED
|
@@ -1,43 +1,72 @@
|
|
| 1 |
{
|
| 2 |
-
|
| 3 |
-
|
| 4 |
-
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
|
| 8 |
-
|
| 9 |
-
|
| 10 |
-
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
|
| 14 |
-
|
| 15 |
-
"context_window_size": 2048,
|
| 16 |
-
"sliding_window_size": 128,
|
| 17 |
-
"prefill_chunk_size": 512,
|
| 18 |
-
"head_dim": 64,
|
| 19 |
-
"dtype": "float32",
|
| 20 |
-
"hidden_act": "silu"
|
| 21 |
-
},
|
| 22 |
"vocab_size": 32768,
|
| 23 |
-
"
|
| 24 |
-
|
| 25 |
-
|
| 26 |
-
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
|
| 30 |
-
|
| 31 |
-
|
| 32 |
-
|
| 33 |
-
|
| 34 |
-
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
|
| 38 |
-
|
| 39 |
-
|
| 40 |
-
|
| 41 |
-
|
| 42 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 43 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"version": "0.1.0",
|
| 3 |
+
"model_type": "qwen2",
|
| 4 |
+
"model_lib": "regular_llm_v1",
|
| 5 |
+
"quantization": "q4f16_1",
|
| 6 |
+
"model_config": {
|
| 7 |
+
"hidden_act": "silu",
|
| 8 |
+
"hidden_size": 768,
|
| 9 |
+
"intermediate_size": 3072,
|
| 10 |
+
"num_attention_heads": 12,
|
| 11 |
+
"num_hidden_layers": 12,
|
| 12 |
+
"num_key_value_heads": 4,
|
| 13 |
+
"rms_norm_eps": 1e-05,
|
| 14 |
+
"rope_theta": 10000.0,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 15 |
"vocab_size": 32768,
|
| 16 |
+
"tie_word_embeddings": true,
|
| 17 |
+
"context_window_size": 2048,
|
| 18 |
+
"prefill_chunk_size": 512,
|
| 19 |
+
"tensor_parallel_shards": 1,
|
| 20 |
+
"head_dim": 64,
|
| 21 |
+
"dtype": "float32"
|
| 22 |
+
},
|
| 23 |
+
"vocab_size": 32768,
|
| 24 |
+
"context_window_size": 2048,
|
| 25 |
+
"sliding_window_size": 128,
|
| 26 |
+
"prefill_chunk_size": 512,
|
| 27 |
+
"attention_sink_size": -1,
|
| 28 |
+
"tensor_parallel_shards": 1,
|
| 29 |
+
"pipeline_parallel_stages": 1,
|
| 30 |
+
"temperature": 0.7,
|
| 31 |
+
"presence_penalty": 0.0,
|
| 32 |
+
"frequency_penalty": 0.0,
|
| 33 |
+
"repetition_penalty": 1.1,
|
| 34 |
+
"top_p": 0.8,
|
| 35 |
+
"tokenizer_files": [
|
| 36 |
+
"tokenizer.json",
|
| 37 |
+
"tokenizer_config.json"
|
| 38 |
+
],
|
| 39 |
+
"tokenizer_info": {
|
| 40 |
+
"token_postproc_method": "byte_level",
|
| 41 |
+
"prepend_space_in_encode": false,
|
| 42 |
+
"strip_space_in_decode": false
|
| 43 |
+
},
|
| 44 |
+
"conv_template": {
|
| 45 |
+
"name": "qwen2",
|
| 46 |
+
"system_template": "<|im_start|>system\n{system_message}<|im_end|>\n",
|
| 47 |
+
"system_message": "You are a helpful AI assistant created by Cihan Yalçın.",
|
| 48 |
+
"roles": {
|
| 49 |
+
"user": "<|im_start|>user",
|
| 50 |
+
"assistant": "<|im_start|>assistant"
|
| 51 |
+
},
|
| 52 |
+
"seps": [
|
| 53 |
+
"<|im_end|>\n"
|
| 54 |
+
],
|
| 55 |
+
"role_content_sep": "\n",
|
| 56 |
+
"role_empty_sep": "\n",
|
| 57 |
+
"stop_str": [
|
| 58 |
+
"<|endoftext|>",
|
| 59 |
+
"<|im_end|>"
|
| 60 |
+
],
|
| 61 |
+
"stop_token_ids": [
|
| 62 |
+
0,
|
| 63 |
+
2
|
| 64 |
+
]
|
| 65 |
+
},
|
| 66 |
+
"pad_token_id": 0,
|
| 67 |
+
"bos_token_id": 0,
|
| 68 |
+
"eos_token_id": [
|
| 69 |
+
2,
|
| 70 |
+
0
|
| 71 |
+
]
|
| 72 |
}
|