Commit
·
c394e13
1
Parent(s):
409299c
增加minicpm
Browse files
MiniCPM-1B-sft-bf16/bert4torch_config.json
ADDED
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token_id": 1,
|
3 |
+
"eos_token_id": 2,
|
4 |
+
"hidden_act": "silu",
|
5 |
+
"hidden_size": 1536,
|
6 |
+
"initializer_range": 0.1,
|
7 |
+
"intermediate_size": 3840,
|
8 |
+
"max_position_embeddings": 4096,
|
9 |
+
"num_attention_heads": 24,
|
10 |
+
"num_hidden_layers": 52,
|
11 |
+
"num_key_value_heads": 8,
|
12 |
+
"layer_norm_eps": 1e-05,
|
13 |
+
"rope_scaling": null,
|
14 |
+
"torch_dtype": "bfloat16",
|
15 |
+
"vocab_size": 73440,
|
16 |
+
"emb_scale": 12,
|
17 |
+
"dim_model_base": 256,
|
18 |
+
"scale_depth": 1.4,
|
19 |
+
"tie_word_embeddings": true,
|
20 |
+
"rope_rank": "updown",
|
21 |
+
"segment_vocab_size": 0,
|
22 |
+
"skip_init": true,
|
23 |
+
"model": "minicpm",
|
24 |
+
"template": "llama3",
|
25 |
+
"convert_lm_logits_dtype": "float32",
|
26 |
+
"generation_config": {"tokenizer_config": {"skip_special_tokens": true, "add_special_tokens": true}, "eos_token_id": [2]}
|
27 |
+
}
|
MiniCPM-2B-128k/bert4torch_config.json
ADDED
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token_id": 1,
|
3 |
+
"eos_token_id": 2,
|
4 |
+
"hidden_act": "silu",
|
5 |
+
"hidden_size": 2304,
|
6 |
+
"initializer_range": 0.1,
|
7 |
+
"intermediate_size": 5760,
|
8 |
+
"max_position_embeddings": 65536,
|
9 |
+
"num_attention_heads": 36,
|
10 |
+
"num_hidden_layers": 40,
|
11 |
+
"num_key_value_heads": 36,
|
12 |
+
"layer_norm_eps": 1e-05,
|
13 |
+
"rope_scaling": {"type": "dynamic", "factor": 4.0},
|
14 |
+
"torch_dtype": "bfloat16",
|
15 |
+
"vocab_size": 122760,
|
16 |
+
"emb_scale": 12,
|
17 |
+
"dim_model_base": 256,
|
18 |
+
"scale_depth": 1.4,
|
19 |
+
"tie_word_embeddings": false,
|
20 |
+
"rope_rank": "updown",
|
21 |
+
"rope_theta": 1000000.0,
|
22 |
+
"segment_vocab_size": 0,
|
23 |
+
"skip_init": true,
|
24 |
+
"model": "minicpm",
|
25 |
+
"template": "llama3",
|
26 |
+
"convert_lm_logits_dtype": "float32",
|
27 |
+
"generation_config": {"tokenizer_config": {"skip_special_tokens": true, "add_special_tokens": true},
|
28 |
+
"eos_token_id": [2], "max_length": 131072}
|
29 |
+
}
|
MiniCPM-2B-dpo-bf16/bert4torch_config.json
ADDED
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token_id": 1,
|
3 |
+
"eos_token_id": 2,
|
4 |
+
"hidden_act": "silu",
|
5 |
+
"hidden_size": 2304,
|
6 |
+
"initializer_range": 0.1,
|
7 |
+
"intermediate_size": 5760,
|
8 |
+
"max_position_embeddings": 4096,
|
9 |
+
"num_attention_heads": 36,
|
10 |
+
"num_hidden_layers": 40,
|
11 |
+
"num_key_value_heads": 36,
|
12 |
+
"layer_norm_eps": 1e-05,
|
13 |
+
"rope_scaling": null,
|
14 |
+
"torch_dtype": "bfloat16",
|
15 |
+
"vocab_size": 122753,
|
16 |
+
"emb_scale": 12,
|
17 |
+
"dim_model_base": 256,
|
18 |
+
"scale_depth": 1.4,
|
19 |
+
"tie_word_embeddings": true,
|
20 |
+
"rope_rank": "updown",
|
21 |
+
"segment_vocab_size": 0,
|
22 |
+
"skip_init": true,
|
23 |
+
"model": "minicpm",
|
24 |
+
"template": "llama3",
|
25 |
+
"convert_lm_logits_dtype": "float32",
|
26 |
+
"generation_config": {"tokenizer_config": {"skip_special_tokens": true, "add_special_tokens": true}, "eos_token_id": [2]}
|
27 |
+
}
|
MiniCPM-2B-sft-bf16/bert4torch_config.json
ADDED
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token_id": 1,
|
3 |
+
"eos_token_id": 2,
|
4 |
+
"hidden_act": "silu",
|
5 |
+
"hidden_size": 2304,
|
6 |
+
"initializer_range": 0.1,
|
7 |
+
"intermediate_size": 5760,
|
8 |
+
"max_position_embeddings": 4096,
|
9 |
+
"num_attention_heads": 36,
|
10 |
+
"num_hidden_layers": 40,
|
11 |
+
"num_key_value_heads": 36,
|
12 |
+
"layer_norm_eps": 1e-05,
|
13 |
+
"rope_scaling": null,
|
14 |
+
"torch_dtype": "bfloat16",
|
15 |
+
"vocab_size": 122753,
|
16 |
+
"emb_scale": 12,
|
17 |
+
"dim_model_base": 256,
|
18 |
+
"scale_depth": 1.4,
|
19 |
+
"tie_word_embeddings": true,
|
20 |
+
"rope_rank": "updown",
|
21 |
+
"segment_vocab_size": 0,
|
22 |
+
"skip_init": true,
|
23 |
+
"model": "minicpm",
|
24 |
+
"template": "llama3",
|
25 |
+
"convert_lm_logits_dtype": "float32",
|
26 |
+
"generation_config": {"tokenizer_config": {"skip_special_tokens": true, "add_special_tokens": true}, "eos_token_id": [2]}
|
27 |
+
}
|