Add files using upload-large-folder tool
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- README.md +8 -0
- chat_template.jinja +14 -0
- config.json +66 -0
- configuration_deepseek.py +199 -0
- generation_config.json +9 -0
- model-00001-of-00088.safetensors +3 -0
- model-00002-of-00088.safetensors +3 -0
- model-00003-of-00088.safetensors +3 -0
- model-00004-of-00088.safetensors +3 -0
- model-00005-of-00088.safetensors +3 -0
- model-00006-of-00088.safetensors +3 -0
- model-00007-of-00088.safetensors +3 -0
- model-00008-of-00088.safetensors +3 -0
- model-00009-of-00088.safetensors +3 -0
- model-00010-of-00088.safetensors +3 -0
- model-00011-of-00088.safetensors +3 -0
- model-00012-of-00088.safetensors +3 -0
- model-00013-of-00088.safetensors +3 -0
- model-00014-of-00088.safetensors +3 -0
- model-00015-of-00088.safetensors +3 -0
- model-00016-of-00088.safetensors +3 -0
- model-00017-of-00088.safetensors +3 -0
- model-00018-of-00088.safetensors +3 -0
- model-00019-of-00088.safetensors +3 -0
- model-00020-of-00088.safetensors +3 -0
- model-00021-of-00088.safetensors +3 -0
- model-00022-of-00088.safetensors +3 -0
- model-00023-of-00088.safetensors +3 -0
- model-00024-of-00088.safetensors +3 -0
- model-00025-of-00088.safetensors +3 -0
- model-00026-of-00088.safetensors +3 -0
- model-00027-of-00088.safetensors +3 -0
- model-00028-of-00088.safetensors +3 -0
- model-00029-of-00088.safetensors +3 -0
- model-00030-of-00088.safetensors +3 -0
- model-00031-of-00088.safetensors +3 -0
- model-00032-of-00088.safetensors +3 -0
- model-00033-of-00088.safetensors +3 -0
- model-00034-of-00088.safetensors +3 -0
- model-00035-of-00088.safetensors +3 -0
- model-00036-of-00088.safetensors +3 -0
- model-00037-of-00088.safetensors +3 -0
- model-00038-of-00088.safetensors +3 -0
- model-00039-of-00088.safetensors +3 -0
- model-00040-of-00088.safetensors +3 -0
- model-00041-of-00088.safetensors +3 -0
- model-00042-of-00088.safetensors +3 -0
- model-00043-of-00088.safetensors +3 -0
- model-00044-of-00088.safetensors +3 -0
- model-00045-of-00088.safetensors +3 -0
README.md
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: mit
|
3 |
+
library_name: mlx
|
4 |
+
base_model: deepseek-ai/DeepSeek-R1-0528
|
5 |
+
tags:
|
6 |
+
- mlx
|
7 |
+
pipeline_tag: text-generation
|
8 |
+
---
|
chat_template.jinja
ADDED
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% set ns = namespace(is_first=false, is_tool=false, is_output_first=true, system_prompt='', is_first_sp=true, is_last_user=false) %}{%- for message in messages %}{%- if message['role'] == 'system' %}{%- if ns.is_first_sp %}{% set ns.system_prompt = ns.system_prompt + message['content'] %}{% set ns.is_first_sp = false %}{%- else %}{% set ns.system_prompt = ns.system_prompt + '
|
2 |
+
|
3 |
+
' + message['content'] %}{%- endif %}{%- endif %}{%- endfor %}{{ bos_token }}{{ ns.system_prompt }}{%- for message in messages %}{% set content = message['content'] %}{%- if message['role'] == 'user' %}{%- set ns.is_tool = false -%}{%- set ns.is_first = false -%}{%- set ns.is_last_user = true -%}{{'<|User|>' + content + '<|Assistant|>'}}{%- endif %}{%- if message['role'] == 'assistant' %}{% if '</think>' in content %}{% set content = content.split('</think>')[-1] %}{% endif %}{% endif %}{%- if message['role'] == 'assistant' and message['tool_calls'] is defined and message['tool_calls'] is not none %}{%- set ns.is_last_user = false -%}{%- if ns.is_tool %}{{'<|tool▁outputs▁end|>'}}{%- endif %}{%- set ns.is_first = false %}{%- set ns.is_tool = false -%}{%- set ns.is_output_first = true %}{%- for tool in message['tool_calls'] %}{%- if not ns.is_first %}{%- if content is none %}{{'<|tool▁calls▁begin|><|tool▁call▁begin|>' + tool['type'] + '<|tool▁sep|>' + tool['function']['name'] + '
|
4 |
+
' + '```json' + '
|
5 |
+
' + tool['function']['arguments'] + '
|
6 |
+
' + '```' + '<|tool▁call▁end|>'}}{%- else %}{{content + '<|tool▁calls▁begin|><|tool▁call▁begin|>' + tool['type'] + '<|tool▁sep|>' + tool['function']['name'] + '
|
7 |
+
' + '```json' + '
|
8 |
+
' + tool['function']['arguments'] + '
|
9 |
+
' + '```' + '<|tool▁call▁end|>'}}{%- endif %}{%- set ns.is_first = true -%}{%- else %}{{'
|
10 |
+
' + '<|tool▁call▁begin|>' + tool['type'] + '<|tool▁sep|>' + tool['function']['name'] + '
|
11 |
+
' + '```json' + '
|
12 |
+
' + tool['function']['arguments'] + '
|
13 |
+
' + '```' + '<|tool▁call▁end|>'}}{%- endif %}{%- endfor %}{{'<|tool▁calls▁end|><|end▁of▁sentence|>'}}{%- endif %}{%- if message['role'] == 'assistant' and (message['tool_calls'] is not defined or message['tool_calls'] is none)%}{%- set ns.is_last_user = false -%}{%- if ns.is_tool %}{{'<|tool▁outputs▁end|>' + content + '<|end▁of▁sentence|>'}}{%- set ns.is_tool = false -%}{%- else %}{{content + '<|end▁of▁sentence|>'}}{%- endif %}{%- endif %}{%- if message['role'] == 'tool' %}{%- set ns.is_last_user = false -%}{%- set ns.is_tool = true -%}{%- if ns.is_output_first %}{{'<|tool▁outputs▁begin|><|tool▁output▁begin|>' + content + '<|tool▁output▁end|>'}}{%- set ns.is_output_first = false %}{%- else %}{{'
|
14 |
+
<|tool▁output▁begin|>' + content + '<|tool▁output▁end|>'}}{%- endif %}{%- endif %}{%- endfor -%}{% if ns.is_tool %}{{'<|tool▁outputs▁end|>'}}{% endif %}{% if add_generation_prompt and not ns.is_last_user and not ns.is_tool %}{{'<|Assistant|>'}}{% endif %}
|
config.json
ADDED
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"architectures": [
|
3 |
+
"DeepseekV3ForCausalLM"
|
4 |
+
],
|
5 |
+
"attention_bias": false,
|
6 |
+
"attention_dropout": 0.0,
|
7 |
+
"auto_map": {
|
8 |
+
"AutoConfig": "configuration_deepseek.DeepseekV3Config",
|
9 |
+
"AutoModel": "modeling_deepseek.DeepseekV3Model",
|
10 |
+
"AutoModelForCausalLM": "modeling_deepseek.DeepseekV3ForCausalLM"
|
11 |
+
},
|
12 |
+
"bos_token_id": 0,
|
13 |
+
"eos_token_id": 1,
|
14 |
+
"ep_size": 1,
|
15 |
+
"first_k_dense_replace": 3,
|
16 |
+
"hidden_act": "silu",
|
17 |
+
"hidden_size": 7168,
|
18 |
+
"initializer_range": 0.02,
|
19 |
+
"intermediate_size": 18432,
|
20 |
+
"kv_lora_rank": 512,
|
21 |
+
"max_position_embeddings": 163840,
|
22 |
+
"model_type": "deepseek_v3",
|
23 |
+
"moe_intermediate_size": 2048,
|
24 |
+
"moe_layer_freq": 1,
|
25 |
+
"n_group": 8,
|
26 |
+
"n_routed_experts": 256,
|
27 |
+
"n_shared_experts": 1,
|
28 |
+
"norm_topk_prob": true,
|
29 |
+
"num_attention_heads": 128,
|
30 |
+
"num_experts_per_tok": 8,
|
31 |
+
"num_hidden_layers": 61,
|
32 |
+
"num_key_value_heads": 128,
|
33 |
+
"num_nextn_predict_layers": 1,
|
34 |
+
"q_lora_rank": 1536,
|
35 |
+
"qk_nope_head_dim": 128,
|
36 |
+
"qk_rope_head_dim": 64,
|
37 |
+
"quantization": {
|
38 |
+
"group_size": 64,
|
39 |
+
"bits": 5
|
40 |
+
},
|
41 |
+
"quantization_config": {
|
42 |
+
"group_size": 64,
|
43 |
+
"bits": 5
|
44 |
+
},
|
45 |
+
"rms_norm_eps": 1e-06,
|
46 |
+
"rope_scaling": {
|
47 |
+
"beta_fast": 32,
|
48 |
+
"beta_slow": 1,
|
49 |
+
"factor": 40,
|
50 |
+
"mscale": 1.0,
|
51 |
+
"mscale_all_dim": 1.0,
|
52 |
+
"original_max_position_embeddings": 4096,
|
53 |
+
"type": "yarn"
|
54 |
+
},
|
55 |
+
"rope_theta": 10000,
|
56 |
+
"routed_scaling_factor": 2.5,
|
57 |
+
"scoring_func": "sigmoid",
|
58 |
+
"tie_word_embeddings": false,
|
59 |
+
"topk_group": 4,
|
60 |
+
"topk_method": "noaux_tc",
|
61 |
+
"torch_dtype": "bfloat16",
|
62 |
+
"transformers_version": "4.46.3",
|
63 |
+
"use_cache": true,
|
64 |
+
"v_head_dim": 128,
|
65 |
+
"vocab_size": 129280
|
66 |
+
}
|
configuration_deepseek.py
ADDED
@@ -0,0 +1,199 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from transformers.configuration_utils import PretrainedConfig
|
2 |
+
from transformers.utils import logging
|
3 |
+
|
4 |
+
logger = logging.get_logger(__name__)
|
5 |
+
|
6 |
+
DEEPSEEK_PRETRAINED_CONFIG_ARCHIVE_MAP = {}
|
7 |
+
class DeepseekV3Config(PretrainedConfig):
|
8 |
+
r"""
|
9 |
+
This is the configuration class to store the configuration of a [`DeepseekV3Model`]. It is used to instantiate an DeepSeek
|
10 |
+
model according to the specified arguments, defining the model architecture. Instantiating a configuration with the
|
11 |
+
defaults will yield a similar configuration to that of the DeepSeek-V3.
|
12 |
+
|
13 |
+
Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
|
14 |
+
documentation from [`PretrainedConfig`] for more information.
|
15 |
+
|
16 |
+
|
17 |
+
Args:
|
18 |
+
vocab_size (`int`, *optional*, defaults to 129280):
|
19 |
+
Vocabulary size of the Deep model. Defines the number of different tokens that can be represented by the
|
20 |
+
`inputs_ids` passed when calling [`DeepseekV3Model`]
|
21 |
+
hidden_size (`int`, *optional*, defaults to 4096):
|
22 |
+
Dimension of the hidden representations.
|
23 |
+
intermediate_size (`int`, *optional*, defaults to 11008):
|
24 |
+
Dimension of the MLP representations.
|
25 |
+
moe_intermediate_size (`int`, *optional*, defaults to 1407):
|
26 |
+
Dimension of the MoE representations.
|
27 |
+
num_hidden_layers (`int`, *optional*, defaults to 32):
|
28 |
+
Number of hidden layers in the Transformer decoder.
|
29 |
+
num_nextn_predict_layers (`int`, *optional*, defaults to 1):
|
30 |
+
Number of nextn predict layers in the DeepSeekV3 Model.
|
31 |
+
num_attention_heads (`int`, *optional*, defaults to 32):
|
32 |
+
Number of attention heads for each attention layer in the Transformer decoder.
|
33 |
+
n_shared_experts (`int`, *optional*, defaults to None):
|
34 |
+
Number of shared experts, None means dense model.
|
35 |
+
n_routed_experts (`int`, *optional*, defaults to None):
|
36 |
+
Number of routed experts, None means dense model.
|
37 |
+
routed_scaling_factor (`float`, *optional*, defaults to 1.0):
|
38 |
+
Scaling factor or routed experts.
|
39 |
+
topk_method (`str`, *optional*, defaults to `gready`):
|
40 |
+
Topk method used in routed gate.
|
41 |
+
n_group (`int`, *optional*, defaults to None):
|
42 |
+
Number of groups for routed experts.
|
43 |
+
topk_group (`int`, *optional*, defaults to None):
|
44 |
+
Number of selected groups for each token(for each token, ensuring the selected experts is only within `topk_group` groups).
|
45 |
+
num_experts_per_tok (`int`, *optional*, defaults to None):
|
46 |
+
Number of selected experts, None means dense model.
|
47 |
+
moe_layer_freq (`int`, *optional*, defaults to 1):
|
48 |
+
The frequency of the MoE layer: one expert layer for every `moe_layer_freq - 1` dense layers.
|
49 |
+
first_k_dense_replace (`int`, *optional*, defaults to 0):
|
50 |
+
Number of dense layers in shallow layers(embed->dense->dense->...->dense->moe->moe...->lm_head).
|
51 |
+
\--k dense layers--/
|
52 |
+
norm_topk_prob (`bool`, *optional*, defaults to False):
|
53 |
+
Whether to normalize the weights of the routed experts.
|
54 |
+
scoring_func (`str`, *optional*, defaults to 'softmax'):
|
55 |
+
Method of computing expert weights.
|
56 |
+
aux_loss_alpha (`float`, *optional*, defaults to 0.001):
|
57 |
+
Auxiliary loss weight coefficient.
|
58 |
+
seq_aux = (`bool`, *optional*, defaults to True):
|
59 |
+
Whether to compute the auxiliary loss for each individual sample.
|
60 |
+
num_key_value_heads (`int`, *optional*):
|
61 |
+
This is the number of key_value heads that should be used to implement Grouped Query Attention. If
|
62 |
+
`num_key_value_heads=num_attention_heads`, the model will use Multi Head Attention (MHA), if
|
63 |
+
`num_key_value_heads=1 the model will use Multi Query Attention (MQA) otherwise GQA is used. When
|
64 |
+
converting a multi-head checkpoint to a GQA checkpoint, each group key and value head should be constructed
|
65 |
+
by meanpooling all the original heads within that group. For more details checkout [this
|
66 |
+
paper](https://arxiv.org/pdf/2305.13245.pdf). If it is not specified, will default to
|
67 |
+
`num_attention_heads`.
|
68 |
+
hidden_act (`str` or `function`, *optional*, defaults to `"silu"`):
|
69 |
+
The non-linear activation function (function or string) in the decoder.
|
70 |
+
max_position_embeddings (`int`, *optional*, defaults to 2048):
|
71 |
+
The maximum sequence length that this model might ever be used with.
|
72 |
+
initializer_range (`float`, *optional*, defaults to 0.02):
|
73 |
+
The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
|
74 |
+
rms_norm_eps (`float`, *optional*, defaults to 1e-06):
|
75 |
+
The epsilon used by the rms normalization layers.
|
76 |
+
use_cache (`bool`, *optional*, defaults to `True`):
|
77 |
+
Whether or not the model should return the last key/values attentions (not used by all models). Only
|
78 |
+
relevant if `config.is_decoder=True`.
|
79 |
+
pad_token_id (`int`, *optional*):
|
80 |
+
Padding token id.
|
81 |
+
bos_token_id (`int`, *optional*, defaults to 1):
|
82 |
+
Beginning of stream token id.
|
83 |
+
eos_token_id (`int`, *optional*, defaults to 2):
|
84 |
+
End of stream token id.
|
85 |
+
tie_word_embeddings (`bool`, *optional*, defaults to `False`):
|
86 |
+
Whether to tie weight embeddings
|
87 |
+
rope_theta (`float`, *optional*, defaults to 10000.0):
|
88 |
+
The base period of the RoPE embeddings.
|
89 |
+
rope_scaling (`Dict`, *optional*):
|
90 |
+
Dictionary containing the scaling configuration for the RoPE embeddings. Currently supports two scaling
|
91 |
+
strategies: linear and dynamic. Their scaling factor must be a float greater than 1. The expected format is
|
92 |
+
`{"type": strategy name, "factor": scaling factor}`. When using this flag, don't update
|
93 |
+
`max_position_embeddings` to the expected new maximum.
|
94 |
+
attention_bias (`bool`, defaults to `False`, *optional*, defaults to `False`):
|
95 |
+
Whether to use a bias in the query, key, value and output projection layers during self-attention.
|
96 |
+
attention_dropout (`float`, *optional*, defaults to 0.0):
|
97 |
+
The dropout ratio for the attention probabilities.
|
98 |
+
|
99 |
+
```python
|
100 |
+
>>> from transformers import DeepseekV3Model, DeepseekV3Config
|
101 |
+
|
102 |
+
>>> # Initializing a Deepseek-V3 style configuration
|
103 |
+
>>> configuration = DeepseekV3Config()
|
104 |
+
|
105 |
+
>>> # Accessing the model configuration
|
106 |
+
>>> configuration = model.config
|
107 |
+
```"""
|
108 |
+
|
109 |
+
model_type = "deepseek_v3"
|
110 |
+
keys_to_ignore_at_inference = ["past_key_values"]
|
111 |
+
|
112 |
+
def __init__(
|
113 |
+
self,
|
114 |
+
vocab_size=129280,
|
115 |
+
hidden_size=7168,
|
116 |
+
intermediate_size=18432,
|
117 |
+
moe_intermediate_size = 2048,
|
118 |
+
num_hidden_layers=61,
|
119 |
+
num_nextn_predict_layers=1,
|
120 |
+
num_attention_heads=128,
|
121 |
+
num_key_value_heads=128,
|
122 |
+
n_shared_experts = 1,
|
123 |
+
n_routed_experts = 256,
|
124 |
+
ep_size = 1,
|
125 |
+
routed_scaling_factor = 2.5,
|
126 |
+
kv_lora_rank = 512,
|
127 |
+
q_lora_rank = 1536,
|
128 |
+
qk_rope_head_dim = 64,
|
129 |
+
v_head_dim = 128,
|
130 |
+
qk_nope_head_dim = 128,
|
131 |
+
topk_method = 'noaux_tc',
|
132 |
+
n_group = 8,
|
133 |
+
topk_group = 4,
|
134 |
+
num_experts_per_tok = 8,
|
135 |
+
moe_layer_freq = 1,
|
136 |
+
first_k_dense_replace = 3,
|
137 |
+
norm_topk_prob = True,
|
138 |
+
scoring_func = 'sigmoid',
|
139 |
+
hidden_act="silu",
|
140 |
+
max_position_embeddings=4096,
|
141 |
+
initializer_range=0.02,
|
142 |
+
rms_norm_eps=1e-6,
|
143 |
+
use_cache=True,
|
144 |
+
pad_token_id=None,
|
145 |
+
bos_token_id=0,
|
146 |
+
eos_token_id=1,
|
147 |
+
tie_word_embeddings=False,
|
148 |
+
rope_theta=10000.0,
|
149 |
+
rope_scaling=None,
|
150 |
+
attention_bias=False,
|
151 |
+
attention_dropout=0.0,
|
152 |
+
**kwargs,
|
153 |
+
):
|
154 |
+
self.vocab_size = vocab_size
|
155 |
+
self.max_position_embeddings = max_position_embeddings
|
156 |
+
self.hidden_size = hidden_size
|
157 |
+
self.intermediate_size = intermediate_size
|
158 |
+
self.moe_intermediate_size = moe_intermediate_size
|
159 |
+
self.num_hidden_layers = num_hidden_layers
|
160 |
+
self.num_nextn_predict_layers = num_nextn_predict_layers
|
161 |
+
self.num_attention_heads = num_attention_heads
|
162 |
+
self.n_shared_experts = n_shared_experts
|
163 |
+
self.n_routed_experts = n_routed_experts
|
164 |
+
self.ep_size = ep_size
|
165 |
+
self.routed_scaling_factor = routed_scaling_factor
|
166 |
+
self.kv_lora_rank = kv_lora_rank
|
167 |
+
self.q_lora_rank = q_lora_rank
|
168 |
+
self.qk_rope_head_dim = qk_rope_head_dim
|
169 |
+
self.v_head_dim = v_head_dim
|
170 |
+
self.qk_nope_head_dim = qk_nope_head_dim
|
171 |
+
self.topk_method = topk_method
|
172 |
+
self.n_group = n_group
|
173 |
+
self.topk_group = topk_group
|
174 |
+
self.num_experts_per_tok = num_experts_per_tok
|
175 |
+
self.moe_layer_freq = moe_layer_freq
|
176 |
+
self.first_k_dense_replace = first_k_dense_replace
|
177 |
+
self.norm_topk_prob = norm_topk_prob
|
178 |
+
self.scoring_func = scoring_func
|
179 |
+
# for backward compatibility
|
180 |
+
if num_key_value_heads is None:
|
181 |
+
num_key_value_heads = num_attention_heads
|
182 |
+
|
183 |
+
self.num_key_value_heads = num_key_value_heads
|
184 |
+
self.hidden_act = hidden_act
|
185 |
+
self.initializer_range = initializer_range
|
186 |
+
self.rms_norm_eps = rms_norm_eps
|
187 |
+
self.use_cache = use_cache
|
188 |
+
self.rope_theta = rope_theta
|
189 |
+
self.rope_scaling = rope_scaling
|
190 |
+
self.attention_bias = attention_bias
|
191 |
+
self.attention_dropout = attention_dropout
|
192 |
+
|
193 |
+
super().__init__(
|
194 |
+
pad_token_id=pad_token_id,
|
195 |
+
bos_token_id=bos_token_id,
|
196 |
+
eos_token_id=eos_token_id,
|
197 |
+
tie_word_embeddings=tie_word_embeddings,
|
198 |
+
**kwargs,
|
199 |
+
)
|
generation_config.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"bos_token_id": 0,
|
4 |
+
"eos_token_id": 1,
|
5 |
+
"do_sample": true,
|
6 |
+
"temperature": 0.6,
|
7 |
+
"top_p": 0.95,
|
8 |
+
"transformers_version": "4.46.3"
|
9 |
+
}
|
model-00001-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:52fcc41b724264e63218547fd9cb6d5970964ddbd0584d3c527c1a00e66df33b
|
3 |
+
size 4552933573
|
model-00002-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6a831265a8fe1499475e915809dbc9a5d45fb5ddbb54ffd499143a715d6086b6
|
3 |
+
size 5330003219
|
model-00003-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b0fe384c9008489ad8c2774f41e731d582e232abc5bfc3fa569e0fa5d187bc92
|
3 |
+
size 5167383320
|
model-00004-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0ef1563a41b679349a8e061e05df949f408d05690c981b44d66ccad867d733b5
|
3 |
+
size 5330003273
|
model-00005-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:33ce25eb58a2c5ea4009e2c4c03a33696065e22d1e0469dec0942c9b8379ae1c
|
3 |
+
size 5330003211
|
model-00006-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:16e51ce2dc127e309230561396a1862f1c1aeba19544dfbe2a447bcafa170a81
|
3 |
+
size 5167383316
|
model-00007-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:810708669726b3f084fc330c1a7e0ef2a9864c4ba855379287b081b1faa1cff3
|
3 |
+
size 5330003271
|
model-00008-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:82297a8aec8304c8206880bfa279f049c4442695fb3ea53eca3f6d18869b9f30
|
3 |
+
size 5330003183
|
model-00009-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cad56e0a7ce35852c18a6ae9f1483de6b8ddff67cde25ed5f21c82570751eba2
|
3 |
+
size 5167383320
|
model-00010-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7726b10ab1fe5ed336eaa59c0343f54364aee1aa855f0fe063d692209aeee91a
|
3 |
+
size 5330003207
|
model-00011-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5b0dff6ed650218c0251f45d498e345a6fbb9c1c145fb318835c8f4038dd6254
|
3 |
+
size 5330003210
|
model-00012-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5393f23063f1609218607d1418be79dc586f33415a0bdaee8b01acee441b0753
|
3 |
+
size 5167383326
|
model-00013-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5db1c7afe0129b685c71569d5d4bfd9467d8c71857516f56ffe5282effcba210
|
3 |
+
size 5330003283
|
model-00014-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:43cad7228ec51ba09fc5915deb0b3efa933c8719776895a2df3a43e84428e9cc
|
3 |
+
size 5330003219
|
model-00015-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9b6c58373146e092fdc26a9ed89da91f7b6b68699bcb5f163b8aa2d5ae932020
|
3 |
+
size 5167383326
|
model-00016-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:ab5d346cdcf45ef699955b618b84fb1b25154f100ce0a700049bf91d54a831c5
|
3 |
+
size 5330003277
|
model-00017-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8577542c3b03a75fd3eed69ca15717e2c90d77a293226ac77f59095d44fe8367
|
3 |
+
size 5330003211
|
model-00018-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f6bfacba26e5414c8b5571790472239d3ac69cd532e8679d760dc4fcd5168500
|
3 |
+
size 5167383322
|
model-00019-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0fced2a91a96180000d1451941381418c15bd241567c088df5cba4cc354d6a3c
|
3 |
+
size 5330003285
|
model-00020-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c4d0eaa18c41fb893b253017d04cb25b996bf07a9c977746b6548e311b302b5d
|
3 |
+
size 5330003245
|
model-00021-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fb24fdd2dd81f9863e64ee088de44bb94abc3497410a8a1b20ceaa2509c16ae8
|
3 |
+
size 5167383326
|
model-00022-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f5af8a230b0d85018ad5ce13c68fa69eeb561405dfbe27b4dff0d36f66144834
|
3 |
+
size 5330003263
|
model-00023-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a7d8011b0fef6ea7b332a24449d6cfdf1dd8a86cf956e1f2b90dae4047fb29d1
|
3 |
+
size 5330003235
|
model-00024-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2622e4f832dddad3eeaaf98ac82f49ad9c75ab94c4abb674cd42cc0e52c02966
|
3 |
+
size 5167383328
|
model-00025-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2455d24b8634f6a25a77162723920d006c6da52cd7e4b8cd6ef4bbefcf361306
|
3 |
+
size 5330003307
|
model-00026-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:15ce152b8c6f85a591ed4545d4ed059b3106cea4f67572721d054c9e6d825222
|
3 |
+
size 5330003229
|
model-00027-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:118b686001f30f8661b4a16727f86b1ab67b7661c44623ced9a7e61ef1ba722f
|
3 |
+
size 5167383322
|
model-00028-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:86f7a8b48426ceba07e6f5cf3bc097d9dd73fb3ee71ab7799c164ad4d8a0af03
|
3 |
+
size 5330003275
|
model-00029-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e355ceb3daa8b923433887ae38a73cbcbb3d10195197c37b500aca7ddde4ae64
|
3 |
+
size 5330003195
|
model-00030-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6a1735ee5ca243f20e45107d9d521417072fb58a1c4daa6f9ec1c0d92553d94d
|
3 |
+
size 5167383324
|
model-00031-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:ab9c3b7b03c87f5d5c02b271eb959df5f564b05a6e8727f3573c48205d8cf940
|
3 |
+
size 5330003309
|
model-00032-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c1ecbb1aeab8c916da47d943d8ae0a39ec0f13b3cf38ac2c6fa8a2e005775f95
|
3 |
+
size 5330003261
|
model-00033-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5849271172f9b44c313492a8c3a1060f6d50a3daee6fbaacc71562615231cf44
|
3 |
+
size 5167383320
|
model-00034-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2e3e1ed02964543d7ffa5c49c4d0710762e67b7fb348ed0935fa32415f02eea7
|
3 |
+
size 5330003263
|
model-00035-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c4c69ab3619c2fae964cfa49c0696f69473de28e55a1fd98f1981d582d580823
|
3 |
+
size 5330003245
|
model-00036-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c3b537b000abd4d50cd27a5e8757595eda059d809e02f3db9385b4a3bf496b66
|
3 |
+
size 5167383326
|
model-00037-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:aa42b38be54f745f3111c46d7b48236b729287ff7a0538dad0ca15d6ea65435a
|
3 |
+
size 5330003309
|
model-00038-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2d6e37fbc83e0fb5b7fc6c859865a31ea229694d33b41742f9819e7e8fb8e003
|
3 |
+
size 5330003221
|
model-00039-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3ea82799d46bf542aac91a6e69c92346fd37bac0cc4c72d11187e975c898d5d9
|
3 |
+
size 5167383322
|
model-00040-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d2f43a9a0aa0345f76ae1eca9281d335a537dc76c5d26d65edc6c21b1c9f0b05
|
3 |
+
size 5330003249
|
model-00041-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2fdae7658cc528ff57f535ca2fa7e15930481b2c961294b9ddc52b5ab010fc65
|
3 |
+
size 5330003225
|
model-00042-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7cec816ec390902710191135f7ac64020af8c4dfa407be705662e9e7a8c87311
|
3 |
+
size 5167383326
|
model-00043-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cd3394b6d47713a707aadeba3c20d16feb58aad4b18265135827b5a750e01b1c
|
3 |
+
size 5330003307
|
model-00044-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0c2848b7717abe00471b7c00a9b5043211bb706e785181316e16519923ef20a3
|
3 |
+
size 5330003251
|
model-00045-of-00088.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:666c71c96ef86927b08a0afbd1a7f550fa537c4d2927ec1c5860585d78cc8309
|
3 |
+
size 5167383326
|