Hugging Face's logo

mlx-community
/
lille-130m-instruct-8bit

Text Generation
MLX
Safetensors
English
lille-130m
conversational
Eval Results
8-bit precision
Model card Files
xet
Community
Goekdeniz-Guelmez commited on
Commit
fb2acaf
·
verified ·
1 Parent(s): db65eca

Add files using upload-large-folder tool

Browse files
Files changed (11) hide show
  1. README.md +125 -0
  2. chat_template.jinja +1 -0
  3. config.json +27 -0
  4. generation_config.json +4 -0
  5. merges.txt +0 -0
  6. model.safetensors +3 -0
  7. model.safetensors.index.json +420 -0
  8. special_tokens_map.json +34 -0
  9. tokenizer.json +0 -0
  10. tokenizer_config.json +59 -0
  11. vocab.json +0 -0
README.md ADDED
@@ -0,0 +1,125 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ datasets:
4
+ - Nikity/Kyoto-Corpus
5
+ language:
6
+ - en
7
+ base_model: mlx-community/lille-130m-instruct-fp16
8
+ base_model_relation: finetune
9
+ pipeline_tag: text-generation
10
+ library_name: mlx
11
+ tags:
12
+ - mlx
13
+ model-index:
14
+ - name: lille-130m-instruct
15
+ results:
16
+ - task:
17
+ type: text-generation
18
+ dataset:
19
+ name: arc_challenge
20
+ type: arc_challenge
21
+ metrics:
22
+ - type: Accuracy
23
+ value: 15.05
24
+ name: ARC (Challenge)
25
+ - task:
26
+ type: text-generation
27
+ dataset:
28
+ name: arc_easy
29
+ type: arc_easy
30
+ metrics:
31
+ - type: Accuracy
32
+ value: 21.4
33
+ name: ARC (Easy)
34
+ - task:
35
+ type: text-generation
36
+ dataset:
37
+ name: gpqa
38
+ type: gpqa
39
+ metrics:
40
+ - type: Accuracy
41
+ value: 12.73
42
+ name: GPQA
43
+ - task:
44
+ type: text-generation
45
+ dataset:
46
+ name: gsm8k
47
+ type: gsm8k
48
+ metrics:
49
+ - type: Accuracy
50
+ value: 7.73
51
+ name: GSM8K
52
+ - task:
53
+ type: text-generation
54
+ dataset:
55
+ name: ifeval
56
+ type: ifeval
57
+ metrics:
58
+ - type: Accuracy
59
+ value: 9.01
60
+ name: IFEVAL
61
+ - task:
62
+ type: text-generation
63
+ dataset:
64
+ name: math
65
+ type: math
66
+ metrics:
67
+ - type: Accuracy
68
+ value: 1.91
69
+ name: MATH (Level 5)
70
+ - task:
71
+ type: text-generation
72
+ dataset:
73
+ name: mmlu
74
+ type: mmlu
75
+ metrics:
76
+ - type: Accuracy
77
+ value: 22.76
78
+ name: MMLU
79
+ - task:
80
+ type: text-generation
81
+ dataset:
82
+ name: mt_bench
83
+ type: mt_bench
84
+ metrics:
85
+ - type: Accuracy
86
+ value: 8.2
87
+ name: MT-Bench
88
+ - task:
89
+ type: text-generation
90
+ dataset:
91
+ name: truthful_qa
92
+ type: truthful_qa
93
+ metrics:
94
+ - type: Accuracy
95
+ value: 9.06
96
+ name: TruthfulQA
97
+ ---
98
+
99
+ # mlx-community/lille-130m-instruct-8bit
100
+
101
+ This model [mlx-community/lille-130m-instruct-8bit](https://huggingface.co/mlx-community/lille-130m-instruct-8bit) was
102
+ converted to MLX format from [mlx-community/lille-130m-instruct-fp16](https://huggingface.co/mlx-community/lille-130m-instruct-fp16)
103
+ using mlx-lm version **0.27.1**.
104
+
105
+ ## Use with mlx
106
+
107
+ ```bash
108
+ pip install mlx-lm
109
+ ```
110
+
111
+ ```python
112
+ from mlx_lm import load, generate
113
+
114
+ model, tokenizer = load("mlx-community/lille-130m-instruct-8bit")
115
+
116
+ prompt = "hello"
117
+
118
+ if tokenizer.chat_template is not None:
119
+ messages = [{"role": "user", "content": prompt}]
120
+ prompt = tokenizer.apply_chat_template(
121
+ messages, add_generation_prompt=True
122
+ )
123
+
124
+ response = generate(model, tokenizer, prompt=prompt, verbose=True)
125
+ ```
chat_template.jinja ADDED
@@ -0,0 +1 @@
 
 
1
+ {{ bos_token }}{% for message in messages %}{% if message['role'] == 'user' %}{{ '<|user|>' + message['content'] + '<|assistant|>' }}{% elif message['role'] == 'assistant' %}{{ message['content'] + eos_token }}{% endif %}{% endfor %}
config.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "LilleForCausalLM"
4
+ ],
5
+ "block_size": 512,
6
+ "dropout": 0.0,
7
+ "layer_norm_eps": 1e-05,
8
+ "model_type": "lille-130m",
9
+ "n_embd": 640,
10
+ "n_head": 10,
11
+ "n_kv_heads": 2,
12
+ "n_layer": 24,
13
+ "quantization": {
14
+ "group_size": 64,
15
+ "bits": 8,
16
+ "mode": "affine"
17
+ },
18
+ "quantization_config": {
19
+ "group_size": 64,
20
+ "bits": 8,
21
+ "mode": "affine"
22
+ },
23
+ "rope_theta": 10000.0,
24
+ "torch_dtype": "float32",
25
+ "transformers_version": "4.53.3",
26
+ "vocab_size": 32768
27
+ }
generation_config.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "transformers_version": "4.53.3"
4
+ }
merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8d258cf51e2079ba9a1472dd5c9b9fddcb84276b776c9f5d8752704c4dd7795a
3
+ size 143206959
model.safetensors.index.json ADDED
@@ -0,0 +1,420 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 143157760,
4
+ "total_parameters": 127171200
5
+ },
6
+ "weight_map": {
7
+ "transformer.layers.0.attention.norm.weight": "model.safetensors",
8
+ "transformer.layers.0.attention.out_proj.biases": "model.safetensors",
9
+ "transformer.layers.0.attention.out_proj.scales": "model.safetensors",
10
+ "transformer.layers.0.attention.out_proj.weight": "model.safetensors",
11
+ "transformer.layers.0.attention.qkv_proj.biases": "model.safetensors",
12
+ "transformer.layers.0.attention.qkv_proj.scales": "model.safetensors",
13
+ "transformer.layers.0.attention.qkv_proj.weight": "model.safetensors",
14
+ "transformer.layers.0.feed_forward.down_proj.biases": "model.safetensors",
15
+ "transformer.layers.0.feed_forward.down_proj.scales": "model.safetensors",
16
+ "transformer.layers.0.feed_forward.down_proj.weight": "model.safetensors",
17
+ "transformer.layers.0.feed_forward.gate_proj.biases": "model.safetensors",
18
+ "transformer.layers.0.feed_forward.gate_proj.scales": "model.safetensors",
19
+ "transformer.layers.0.feed_forward.gate_proj.weight": "model.safetensors",
20
+ "transformer.layers.0.feed_forward.norm.weight": "model.safetensors",
21
+ "transformer.layers.0.feed_forward.up_proj.biases": "model.safetensors",
22
+ "transformer.layers.0.feed_forward.up_proj.scales": "model.safetensors",
23
+ "transformer.layers.0.feed_forward.up_proj.weight": "model.safetensors",
24
+ "transformer.layers.1.attention.norm.weight": "model.safetensors",
25
+ "transformer.layers.1.attention.out_proj.biases": "model.safetensors",
26
+ "transformer.layers.1.attention.out_proj.scales": "model.safetensors",
27
+ "transformer.layers.1.attention.out_proj.weight": "model.safetensors",
28
+ "transformer.layers.1.attention.qkv_proj.biases": "model.safetensors",
29
+ "transformer.layers.1.attention.qkv_proj.scales": "model.safetensors",
30
+ "transformer.layers.1.attention.qkv_proj.weight": "model.safetensors",
31
+ "transformer.layers.1.feed_forward.down_proj.biases": "model.safetensors",
32
+ "transformer.layers.1.feed_forward.down_proj.scales": "model.safetensors",
33
+ "transformer.layers.1.feed_forward.down_proj.weight": "model.safetensors",
34
+ "transformer.layers.1.feed_forward.gate_proj.biases": "model.safetensors",
35
+ "transformer.layers.1.feed_forward.gate_proj.scales": "model.safetensors",
36
+ "transformer.layers.1.feed_forward.gate_proj.weight": "model.safetensors",
37
+ "transformer.layers.1.feed_forward.norm.weight": "model.safetensors",
38
+ "transformer.layers.1.feed_forward.up_proj.biases": "model.safetensors",
39
+ "transformer.layers.1.feed_forward.up_proj.scales": "model.safetensors",
40
+ "transformer.layers.1.feed_forward.up_proj.weight": "model.safetensors",
41
+ "transformer.layers.10.attention.norm.weight": "model.safetensors",
42
+ "transformer.layers.10.attention.out_proj.biases": "model.safetensors",
43
+ "transformer.layers.10.attention.out_proj.scales": "model.safetensors",
44
+ "transformer.layers.10.attention.out_proj.weight": "model.safetensors",
45
+ "transformer.layers.10.attention.qkv_proj.biases": "model.safetensors",
46
+ "transformer.layers.10.attention.qkv_proj.scales": "model.safetensors",
47
+ "transformer.layers.10.attention.qkv_proj.weight": "model.safetensors",
48
+ "transformer.layers.10.feed_forward.down_proj.biases": "model.safetensors",
49
+ "transformer.layers.10.feed_forward.down_proj.scales": "model.safetensors",
50
+ "transformer.layers.10.feed_forward.down_proj.weight": "model.safetensors",
51
+ "transformer.layers.10.feed_forward.gate_proj.biases": "model.safetensors",
52
+ "transformer.layers.10.feed_forward.gate_proj.scales": "model.safetensors",
53
+ "transformer.layers.10.feed_forward.gate_proj.weight": "model.safetensors",
54
+ "transformer.layers.10.feed_forward.norm.weight": "model.safetensors",
55
+ "transformer.layers.10.feed_forward.up_proj.biases": "model.safetensors",
56
+ "transformer.layers.10.feed_forward.up_proj.scales": "model.safetensors",
57
+ "transformer.layers.10.feed_forward.up_proj.weight": "model.safetensors",
58
+ "transformer.layers.11.attention.norm.weight": "model.safetensors",
59
+ "transformer.layers.11.attention.out_proj.biases": "model.safetensors",
60
+ "transformer.layers.11.attention.out_proj.scales": "model.safetensors",
61
+ "transformer.layers.11.attention.out_proj.weight": "model.safetensors",
62
+ "transformer.layers.11.attention.qkv_proj.biases": "model.safetensors",
63
+ "transformer.layers.11.attention.qkv_proj.scales": "model.safetensors",
64
+ "transformer.layers.11.attention.qkv_proj.weight": "model.safetensors",
65
+ "transformer.layers.11.feed_forward.down_proj.biases": "model.safetensors",
66
+ "transformer.layers.11.feed_forward.down_proj.scales": "model.safetensors",
67
+ "transformer.layers.11.feed_forward.down_proj.weight": "model.safetensors",
68
+ "transformer.layers.11.feed_forward.gate_proj.biases": "model.safetensors",
69
+ "transformer.layers.11.feed_forward.gate_proj.scales": "model.safetensors",
70
+ "transformer.layers.11.feed_forward.gate_proj.weight": "model.safetensors",
71
+ "transformer.layers.11.feed_forward.norm.weight": "model.safetensors",
72
+ "transformer.layers.11.feed_forward.up_proj.biases": "model.safetensors",
73
+ "transformer.layers.11.feed_forward.up_proj.scales": "model.safetensors",
74
+ "transformer.layers.11.feed_forward.up_proj.weight": "model.safetensors",
75
+ "transformer.layers.12.attention.norm.weight": "model.safetensors",
76
+ "transformer.layers.12.attention.out_proj.biases": "model.safetensors",
77
+ "transformer.layers.12.attention.out_proj.scales": "model.safetensors",
78
+ "transformer.layers.12.attention.out_proj.weight": "model.safetensors",
79
+ "transformer.layers.12.attention.qkv_proj.biases": "model.safetensors",
80
+ "transformer.layers.12.attention.qkv_proj.scales": "model.safetensors",
81
+ "transformer.layers.12.attention.qkv_proj.weight": "model.safetensors",
82
+ "transformer.layers.12.feed_forward.down_proj.biases": "model.safetensors",
83
+ "transformer.layers.12.feed_forward.down_proj.scales": "model.safetensors",
84
+ "transformer.layers.12.feed_forward.down_proj.weight": "model.safetensors",
85
+ "transformer.layers.12.feed_forward.gate_proj.biases": "model.safetensors",
86
+ "transformer.layers.12.feed_forward.gate_proj.scales": "model.safetensors",
87
+ "transformer.layers.12.feed_forward.gate_proj.weight": "model.safetensors",
88
+ "transformer.layers.12.feed_forward.norm.weight": "model.safetensors",
89
+ "transformer.layers.12.feed_forward.up_proj.biases": "model.safetensors",
90
+ "transformer.layers.12.feed_forward.up_proj.scales": "model.safetensors",
91
+ "transformer.layers.12.feed_forward.up_proj.weight": "model.safetensors",
92
+ "transformer.layers.13.attention.norm.weight": "model.safetensors",
93
+ "transformer.layers.13.attention.out_proj.biases": "model.safetensors",
94
+ "transformer.layers.13.attention.out_proj.scales": "model.safetensors",
95
+ "transformer.layers.13.attention.out_proj.weight": "model.safetensors",
96
+ "transformer.layers.13.attention.qkv_proj.biases": "model.safetensors",
97
+ "transformer.layers.13.attention.qkv_proj.scales": "model.safetensors",
98
+ "transformer.layers.13.attention.qkv_proj.weight": "model.safetensors",
99
+ "transformer.layers.13.feed_forward.down_proj.biases": "model.safetensors",
100
+ "transformer.layers.13.feed_forward.down_proj.scales": "model.safetensors",
101
+ "transformer.layers.13.feed_forward.down_proj.weight": "model.safetensors",
102
+ "transformer.layers.13.feed_forward.gate_proj.biases": "model.safetensors",
103
+ "transformer.layers.13.feed_forward.gate_proj.scales": "model.safetensors",
104
+ "transformer.layers.13.feed_forward.gate_proj.weight": "model.safetensors",
105
+ "transformer.layers.13.feed_forward.norm.weight": "model.safetensors",
106
+ "transformer.layers.13.feed_forward.up_proj.biases": "model.safetensors",
107
+ "transformer.layers.13.feed_forward.up_proj.scales": "model.safetensors",
108
+ "transformer.layers.13.feed_forward.up_proj.weight": "model.safetensors",
109
+ "transformer.layers.14.attention.norm.weight": "model.safetensors",
110
+ "transformer.layers.14.attention.out_proj.biases": "model.safetensors",
111
+ "transformer.layers.14.attention.out_proj.scales": "model.safetensors",
112
+ "transformer.layers.14.attention.out_proj.weight": "model.safetensors",
113
+ "transformer.layers.14.attention.qkv_proj.biases": "model.safetensors",
114
+ "transformer.layers.14.attention.qkv_proj.scales": "model.safetensors",
115
+ "transformer.layers.14.attention.qkv_proj.weight": "model.safetensors",
116
+ "transformer.layers.14.feed_forward.down_proj.biases": "model.safetensors",
117
+ "transformer.layers.14.feed_forward.down_proj.scales": "model.safetensors",
118
+ "transformer.layers.14.feed_forward.down_proj.weight": "model.safetensors",
119
+ "transformer.layers.14.feed_forward.gate_proj.biases": "model.safetensors",
120
+ "transformer.layers.14.feed_forward.gate_proj.scales": "model.safetensors",
121
+ "transformer.layers.14.feed_forward.gate_proj.weight": "model.safetensors",
122
+ "transformer.layers.14.feed_forward.norm.weight": "model.safetensors",
123
+ "transformer.layers.14.feed_forward.up_proj.biases": "model.safetensors",
124
+ "transformer.layers.14.feed_forward.up_proj.scales": "model.safetensors",
125
+ "transformer.layers.14.feed_forward.up_proj.weight": "model.safetensors",
126
+ "transformer.layers.15.attention.norm.weight": "model.safetensors",
127
+ "transformer.layers.15.attention.out_proj.biases": "model.safetensors",
128
+ "transformer.layers.15.attention.out_proj.scales": "model.safetensors",
129
+ "transformer.layers.15.attention.out_proj.weight": "model.safetensors",
130
+ "transformer.layers.15.attention.qkv_proj.biases": "model.safetensors",
131
+ "transformer.layers.15.attention.qkv_proj.scales": "model.safetensors",
132
+ "transformer.layers.15.attention.qkv_proj.weight": "model.safetensors",
133
+ "transformer.layers.15.feed_forward.down_proj.biases": "model.safetensors",
134
+ "transformer.layers.15.feed_forward.down_proj.scales": "model.safetensors",
135
+ "transformer.layers.15.feed_forward.down_proj.weight": "model.safetensors",
136
+ "transformer.layers.15.feed_forward.gate_proj.biases": "model.safetensors",
137
+ "transformer.layers.15.feed_forward.gate_proj.scales": "model.safetensors",
138
+ "transformer.layers.15.feed_forward.gate_proj.weight": "model.safetensors",
139
+ "transformer.layers.15.feed_forward.norm.weight": "model.safetensors",
140
+ "transformer.layers.15.feed_forward.up_proj.biases": "model.safetensors",
141
+ "transformer.layers.15.feed_forward.up_proj.scales": "model.safetensors",
142
+ "transformer.layers.15.feed_forward.up_proj.weight": "model.safetensors",
143
+ "transformer.layers.16.attention.norm.weight": "model.safetensors",
144
+ "transformer.layers.16.attention.out_proj.biases": "model.safetensors",
145
+ "transformer.layers.16.attention.out_proj.scales": "model.safetensors",
146
+ "transformer.layers.16.attention.out_proj.weight": "model.safetensors",
147
+ "transformer.layers.16.attention.qkv_proj.biases": "model.safetensors",
148
+ "transformer.layers.16.attention.qkv_proj.scales": "model.safetensors",
149
+ "transformer.layers.16.attention.qkv_proj.weight": "model.safetensors",
150
+ "transformer.layers.16.feed_forward.down_proj.biases": "model.safetensors",
151
+ "transformer.layers.16.feed_forward.down_proj.scales": "model.safetensors",
152
+ "transformer.layers.16.feed_forward.down_proj.weight": "model.safetensors",
153
+ "transformer.layers.16.feed_forward.gate_proj.biases": "model.safetensors",
154
+ "transformer.layers.16.feed_forward.gate_proj.scales": "model.safetensors",
155
+ "transformer.layers.16.feed_forward.gate_proj.weight": "model.safetensors",
156
+ "transformer.layers.16.feed_forward.norm.weight": "model.safetensors",
157
+ "transformer.layers.16.feed_forward.up_proj.biases": "model.safetensors",
158
+ "transformer.layers.16.feed_forward.up_proj.scales": "model.safetensors",
159
+ "transformer.layers.16.feed_forward.up_proj.weight": "model.safetensors",
160
+ "transformer.layers.17.attention.norm.weight": "model.safetensors",
161
+ "transformer.layers.17.attention.out_proj.biases": "model.safetensors",
162
+ "transformer.layers.17.attention.out_proj.scales": "model.safetensors",
163
+ "transformer.layers.17.attention.out_proj.weight": "model.safetensors",
164
+ "transformer.layers.17.attention.qkv_proj.biases": "model.safetensors",
165
+ "transformer.layers.17.attention.qkv_proj.scales": "model.safetensors",
166
+ "transformer.layers.17.attention.qkv_proj.weight": "model.safetensors",
167
+ "transformer.layers.17.feed_forward.down_proj.biases": "model.safetensors",
168
+ "transformer.layers.17.feed_forward.down_proj.scales": "model.safetensors",
169
+ "transformer.layers.17.feed_forward.down_proj.weight": "model.safetensors",
170
+ "transformer.layers.17.feed_forward.gate_proj.biases": "model.safetensors",
171
+ "transformer.layers.17.feed_forward.gate_proj.scales": "model.safetensors",
172
+ "transformer.layers.17.feed_forward.gate_proj.weight": "model.safetensors",
173
+ "transformer.layers.17.feed_forward.norm.weight": "model.safetensors",
174
+ "transformer.layers.17.feed_forward.up_proj.biases": "model.safetensors",
175
+ "transformer.layers.17.feed_forward.up_proj.scales": "model.safetensors",
176
+ "transformer.layers.17.feed_forward.up_proj.weight": "model.safetensors",
177
+ "transformer.layers.18.attention.norm.weight": "model.safetensors",
178
+ "transformer.layers.18.attention.out_proj.biases": "model.safetensors",
179
+ "transformer.layers.18.attention.out_proj.scales": "model.safetensors",
180
+ "transformer.layers.18.attention.out_proj.weight": "model.safetensors",
181
+ "transformer.layers.18.attention.qkv_proj.biases": "model.safetensors",
182
+ "transformer.layers.18.attention.qkv_proj.scales": "model.safetensors",
183
+ "transformer.layers.18.attention.qkv_proj.weight": "model.safetensors",
184
+ "transformer.layers.18.feed_forward.down_proj.biases": "model.safetensors",
185
+ "transformer.layers.18.feed_forward.down_proj.scales": "model.safetensors",
186
+ "transformer.layers.18.feed_forward.down_proj.weight": "model.safetensors",
187
+ "transformer.layers.18.feed_forward.gate_proj.biases": "model.safetensors",
188
+ "transformer.layers.18.feed_forward.gate_proj.scales": "model.safetensors",
189
+ "transformer.layers.18.feed_forward.gate_proj.weight": "model.safetensors",
190
+ "transformer.layers.18.feed_forward.norm.weight": "model.safetensors",
191
+ "transformer.layers.18.feed_forward.up_proj.biases": "model.safetensors",
192
+ "transformer.layers.18.feed_forward.up_proj.scales": "model.safetensors",
193
+ "transformer.layers.18.feed_forward.up_proj.weight": "model.safetensors",
194
+ "transformer.layers.19.attention.norm.weight": "model.safetensors",
195
+ "transformer.layers.19.attention.out_proj.biases": "model.safetensors",
196
+ "transformer.layers.19.attention.out_proj.scales": "model.safetensors",
197
+ "transformer.layers.19.attention.out_proj.weight": "model.safetensors",
198
+ "transformer.layers.19.attention.qkv_proj.biases": "model.safetensors",
199
+ "transformer.layers.19.attention.qkv_proj.scales": "model.safetensors",
200
+ "transformer.layers.19.attention.qkv_proj.weight": "model.safetensors",
201
+ "transformer.layers.19.feed_forward.down_proj.biases": "model.safetensors",
202
+ "transformer.layers.19.feed_forward.down_proj.scales": "model.safetensors",
203
+ "transformer.layers.19.feed_forward.down_proj.weight": "model.safetensors",
204
+ "transformer.layers.19.feed_forward.gate_proj.biases": "model.safetensors",
205
+ "transformer.layers.19.feed_forward.gate_proj.scales": "model.safetensors",
206
+ "transformer.layers.19.feed_forward.gate_proj.weight": "model.safetensors",
207
+ "transformer.layers.19.feed_forward.norm.weight": "model.safetensors",
208
+ "transformer.layers.19.feed_forward.up_proj.biases": "model.safetensors",
209
+ "transformer.layers.19.feed_forward.up_proj.scales": "model.safetensors",
210
+ "transformer.layers.19.feed_forward.up_proj.weight": "model.safetensors",
211
+ "transformer.layers.2.attention.norm.weight": "model.safetensors",
212
+ "transformer.layers.2.attention.out_proj.biases": "model.safetensors",
213
+ "transformer.layers.2.attention.out_proj.scales": "model.safetensors",
214
+ "transformer.layers.2.attention.out_proj.weight": "model.safetensors",
215
+ "transformer.layers.2.attention.qkv_proj.biases": "model.safetensors",
216
+ "transformer.layers.2.attention.qkv_proj.scales": "model.safetensors",
217
+ "transformer.layers.2.attention.qkv_proj.weight": "model.safetensors",
218
+ "transformer.layers.2.feed_forward.down_proj.biases": "model.safetensors",
219
+ "transformer.layers.2.feed_forward.down_proj.scales": "model.safetensors",
220
+ "transformer.layers.2.feed_forward.down_proj.weight": "model.safetensors",
221
+ "transformer.layers.2.feed_forward.gate_proj.biases": "model.safetensors",
222
+ "transformer.layers.2.feed_forward.gate_proj.scales": "model.safetensors",
223
+ "transformer.layers.2.feed_forward.gate_proj.weight": "model.safetensors",
224
+ "transformer.layers.2.feed_forward.norm.weight": "model.safetensors",
225
+ "transformer.layers.2.feed_forward.up_proj.biases": "model.safetensors",
226
+ "transformer.layers.2.feed_forward.up_proj.scales": "model.safetensors",
227
+ "transformer.layers.2.feed_forward.up_proj.weight": "model.safetensors",
228
+ "transformer.layers.20.attention.norm.weight": "model.safetensors",
229
+ "transformer.layers.20.attention.out_proj.biases": "model.safetensors",
230
+ "transformer.layers.20.attention.out_proj.scales": "model.safetensors",
231
+ "transformer.layers.20.attention.out_proj.weight": "model.safetensors",
232
+ "transformer.layers.20.attention.qkv_proj.biases": "model.safetensors",
233
+ "transformer.layers.20.attention.qkv_proj.scales": "model.safetensors",
234
+ "transformer.layers.20.attention.qkv_proj.weight": "model.safetensors",
235
+ "transformer.layers.20.feed_forward.down_proj.biases": "model.safetensors",
236
+ "transformer.layers.20.feed_forward.down_proj.scales": "model.safetensors",
237
+ "transformer.layers.20.feed_forward.down_proj.weight": "model.safetensors",
238
+ "transformer.layers.20.feed_forward.gate_proj.biases": "model.safetensors",
239
+ "transformer.layers.20.feed_forward.gate_proj.scales": "model.safetensors",
240
+ "transformer.layers.20.feed_forward.gate_proj.weight": "model.safetensors",
241
+ "transformer.layers.20.feed_forward.norm.weight": "model.safetensors",
242
+ "transformer.layers.20.feed_forward.up_proj.biases": "model.safetensors",
243
+ "transformer.layers.20.feed_forward.up_proj.scales": "model.safetensors",
244
+ "transformer.layers.20.feed_forward.up_proj.weight": "model.safetensors",
245
+ "transformer.layers.21.attention.norm.weight": "model.safetensors",
246
+ "transformer.layers.21.attention.out_proj.biases": "model.safetensors",
247
+ "transformer.layers.21.attention.out_proj.scales": "model.safetensors",
248
+ "transformer.layers.21.attention.out_proj.weight": "model.safetensors",
249
+ "transformer.layers.21.attention.qkv_proj.biases": "model.safetensors",
250
+ "transformer.layers.21.attention.qkv_proj.scales": "model.safetensors",
251
+ "transformer.layers.21.attention.qkv_proj.weight": "model.safetensors",
252
+ "transformer.layers.21.feed_forward.down_proj.biases": "model.safetensors",
253
+ "transformer.layers.21.feed_forward.down_proj.scales": "model.safetensors",
254
+ "transformer.layers.21.feed_forward.down_proj.weight": "model.safetensors",
255
+ "transformer.layers.21.feed_forward.gate_proj.biases": "model.safetensors",
256
+ "transformer.layers.21.feed_forward.gate_proj.scales": "model.safetensors",
257
+ "transformer.layers.21.feed_forward.gate_proj.weight": "model.safetensors",
258
+ "transformer.layers.21.feed_forward.norm.weight": "model.safetensors",
259
+ "transformer.layers.21.feed_forward.up_proj.biases": "model.safetensors",
260
+ "transformer.layers.21.feed_forward.up_proj.scales": "model.safetensors",
261
+ "transformer.layers.21.feed_forward.up_proj.weight": "model.safetensors",
262
+ "transformer.layers.22.attention.norm.weight": "model.safetensors",
263
+ "transformer.layers.22.attention.out_proj.biases": "model.safetensors",
264
+ "transformer.layers.22.attention.out_proj.scales": "model.safetensors",
265
+ "transformer.layers.22.attention.out_proj.weight": "model.safetensors",
266
+ "transformer.layers.22.attention.qkv_proj.biases": "model.safetensors",
267
+ "transformer.layers.22.attention.qkv_proj.scales": "model.safetensors",
268
+ "transformer.layers.22.attention.qkv_proj.weight": "model.safetensors",
269
+ "transformer.layers.22.feed_forward.down_proj.biases": "model.safetensors",
270
+ "transformer.layers.22.feed_forward.down_proj.scales": "model.safetensors",
271
+ "transformer.layers.22.feed_forward.down_proj.weight": "model.safetensors",
272
+ "transformer.layers.22.feed_forward.gate_proj.biases": "model.safetensors",
273
+ "transformer.layers.22.feed_forward.gate_proj.scales": "model.safetensors",
274
+ "transformer.layers.22.feed_forward.gate_proj.weight": "model.safetensors",
275
+ "transformer.layers.22.feed_forward.norm.weight": "model.safetensors",
276
+ "transformer.layers.22.feed_forward.up_proj.biases": "model.safetensors",
277
+ "transformer.layers.22.feed_forward.up_proj.scales": "model.safetensors",
278
+ "transformer.layers.22.feed_forward.up_proj.weight": "model.safetensors",
279
+ "transformer.layers.23.attention.norm.weight": "model.safetensors",
280
+ "transformer.layers.23.attention.out_proj.biases": "model.safetensors",
281
+ "transformer.layers.23.attention.out_proj.scales": "model.safetensors",
282
+ "transformer.layers.23.attention.out_proj.weight": "model.safetensors",
283
+ "transformer.layers.23.attention.qkv_proj.biases": "model.safetensors",
284
+ "transformer.layers.23.attention.qkv_proj.scales": "model.safetensors",
285
+ "transformer.layers.23.attention.qkv_proj.weight": "model.safetensors",
286
+ "transformer.layers.23.feed_forward.down_proj.biases": "model.safetensors",
287
+ "transformer.layers.23.feed_forward.down_proj.scales": "model.safetensors",
288
+ "transformer.layers.23.feed_forward.down_proj.weight": "model.safetensors",
289
+ "transformer.layers.23.feed_forward.gate_proj.biases": "model.safetensors",
290
+ "transformer.layers.23.feed_forward.gate_proj.scales": "model.safetensors",
291
+ "transformer.layers.23.feed_forward.gate_proj.weight": "model.safetensors",
292
+ "transformer.layers.23.feed_forward.norm.weight": "model.safetensors",
293
+ "transformer.layers.23.feed_forward.up_proj.biases": "model.safetensors",
294
+ "transformer.layers.23.feed_forward.up_proj.scales": "model.safetensors",
295
+ "transformer.layers.23.feed_forward.up_proj.weight": "model.safetensors",
296
+ "transformer.layers.3.attention.norm.weight": "model.safetensors",
297
+ "transformer.layers.3.attention.out_proj.biases": "model.safetensors",
298
+ "transformer.layers.3.attention.out_proj.scales": "model.safetensors",
299
+ "transformer.layers.3.attention.out_proj.weight": "model.safetensors",
300
+ "transformer.layers.3.attention.qkv_proj.biases": "model.safetensors",
301
+ "transformer.layers.3.attention.qkv_proj.scales": "model.safetensors",
302
+ "transformer.layers.3.attention.qkv_proj.weight": "model.safetensors",
303
+ "transformer.layers.3.feed_forward.down_proj.biases": "model.safetensors",
304
+ "transformer.layers.3.feed_forward.down_proj.scales": "model.safetensors",
305
+ "transformer.layers.3.feed_forward.down_proj.weight": "model.safetensors",
306
+ "transformer.layers.3.feed_forward.gate_proj.biases": "model.safetensors",
307
+ "transformer.layers.3.feed_forward.gate_proj.scales": "model.safetensors",
308
+ "transformer.layers.3.feed_forward.gate_proj.weight": "model.safetensors",
309
+ "transformer.layers.3.feed_forward.norm.weight": "model.safetensors",
310
+ "transformer.layers.3.feed_forward.up_proj.biases": "model.safetensors",
311
+ "transformer.layers.3.feed_forward.up_proj.scales": "model.safetensors",
312
+ "transformer.layers.3.feed_forward.up_proj.weight": "model.safetensors",
313
+ "transformer.layers.4.attention.norm.weight": "model.safetensors",
314
+ "transformer.layers.4.attention.out_proj.biases": "model.safetensors",
315
+ "transformer.layers.4.attention.out_proj.scales": "model.safetensors",
316
+ "transformer.layers.4.attention.out_proj.weight": "model.safetensors",
317
+ "transformer.layers.4.attention.qkv_proj.biases": "model.safetensors",
318
+ "transformer.layers.4.attention.qkv_proj.scales": "model.safetensors",
319
+ "transformer.layers.4.attention.qkv_proj.weight": "model.safetensors",
320
+ "transformer.layers.4.feed_forward.down_proj.biases": "model.safetensors",
321
+ "transformer.layers.4.feed_forward.down_proj.scales": "model.safetensors",
322
+ "transformer.layers.4.feed_forward.down_proj.weight": "model.safetensors",
323
+ "transformer.layers.4.feed_forward.gate_proj.biases": "model.safetensors",
324
+ "transformer.layers.4.feed_forward.gate_proj.scales": "model.safetensors",
325
+ "transformer.layers.4.feed_forward.gate_proj.weight": "model.safetensors",
326
+ "transformer.layers.4.feed_forward.norm.weight": "model.safetensors",
327
+ "transformer.layers.4.feed_forward.up_proj.biases": "model.safetensors",
328
+ "transformer.layers.4.feed_forward.up_proj.scales": "model.safetensors",
329
+ "transformer.layers.4.feed_forward.up_proj.weight": "model.safetensors",
330
+ "transformer.layers.5.attention.norm.weight": "model.safetensors",
331
+ "transformer.layers.5.attention.out_proj.biases": "model.safetensors",
332
+ "transformer.layers.5.attention.out_proj.scales": "model.safetensors",
333
+ "transformer.layers.5.attention.out_proj.weight": "model.safetensors",
334
+ "transformer.layers.5.attention.qkv_proj.biases": "model.safetensors",
335
+ "transformer.layers.5.attention.qkv_proj.scales": "model.safetensors",
336
+ "transformer.layers.5.attention.qkv_proj.weight": "model.safetensors",
337
+ "transformer.layers.5.feed_forward.down_proj.biases": "model.safetensors",
338
+ "transformer.layers.5.feed_forward.down_proj.scales": "model.safetensors",
339
+ "transformer.layers.5.feed_forward.down_proj.weight": "model.safetensors",
340
+ "transformer.layers.5.feed_forward.gate_proj.biases": "model.safetensors",
341
+ "transformer.layers.5.feed_forward.gate_proj.scales": "model.safetensors",
342
+ "transformer.layers.5.feed_forward.gate_proj.weight": "model.safetensors",
343
+ "transformer.layers.5.feed_forward.norm.weight": "model.safetensors",
344
+ "transformer.layers.5.feed_forward.up_proj.biases": "model.safetensors",
345
+ "transformer.layers.5.feed_forward.up_proj.scales": "model.safetensors",
346
+ "transformer.layers.5.feed_forward.up_proj.weight": "model.safetensors",
347
+ "transformer.layers.6.attention.norm.weight": "model.safetensors",
348
+ "transformer.layers.6.attention.out_proj.biases": "model.safetensors",
349
+ "transformer.layers.6.attention.out_proj.scales": "model.safetensors",
350
+ "transformer.layers.6.attention.out_proj.weight": "model.safetensors",
351
+ "transformer.layers.6.attention.qkv_proj.biases": "model.safetensors",
352
+ "transformer.layers.6.attention.qkv_proj.scales": "model.safetensors",
353
+ "transformer.layers.6.attention.qkv_proj.weight": "model.safetensors",
354
+ "transformer.layers.6.feed_forward.down_proj.biases": "model.safetensors",
355
+ "transformer.layers.6.feed_forward.down_proj.scales": "model.safetensors",
356
+ "transformer.layers.6.feed_forward.down_proj.weight": "model.safetensors",
357
+ "transformer.layers.6.feed_forward.gate_proj.biases": "model.safetensors",
358
+ "transformer.layers.6.feed_forward.gate_proj.scales": "model.safetensors",
359
+ "transformer.layers.6.feed_forward.gate_proj.weight": "model.safetensors",
360
+ "transformer.layers.6.feed_forward.norm.weight": "model.safetensors",
361
+ "transformer.layers.6.feed_forward.up_proj.biases": "model.safetensors",
362
+ "transformer.layers.6.feed_forward.up_proj.scales": "model.safetensors",
363
+ "transformer.layers.6.feed_forward.up_proj.weight": "model.safetensors",
364
+ "transformer.layers.7.attention.norm.weight": "model.safetensors",
365
+ "transformer.layers.7.attention.out_proj.biases": "model.safetensors",
366
+ "transformer.layers.7.attention.out_proj.scales": "model.safetensors",
367
+ "transformer.layers.7.attention.out_proj.weight": "model.safetensors",
368
+ "transformer.layers.7.attention.qkv_proj.biases": "model.safetensors",
369
+ "transformer.layers.7.attention.qkv_proj.scales": "model.safetensors",
370
+ "transformer.layers.7.attention.qkv_proj.weight": "model.safetensors",
371
+ "transformer.layers.7.feed_forward.down_proj.biases": "model.safetensors",
372
+ "transformer.layers.7.feed_forward.down_proj.scales": "model.safetensors",
373
+ "transformer.layers.7.feed_forward.down_proj.weight": "model.safetensors",
374
+ "transformer.layers.7.feed_forward.gate_proj.biases": "model.safetensors",
375
+ "transformer.layers.7.feed_forward.gate_proj.scales": "model.safetensors",
376
+ "transformer.layers.7.feed_forward.gate_proj.weight": "model.safetensors",
377
+ "transformer.layers.7.feed_forward.norm.weight": "model.safetensors",
378
+ "transformer.layers.7.feed_forward.up_proj.biases": "model.safetensors",
379
+ "transformer.layers.7.feed_forward.up_proj.scales": "model.safetensors",
380
+ "transformer.layers.7.feed_forward.up_proj.weight": "model.safetensors",
381
+ "transformer.layers.8.attention.norm.weight": "model.safetensors",
382
+ "transformer.layers.8.attention.out_proj.biases": "model.safetensors",
383
+ "transformer.layers.8.attention.out_proj.scales": "model.safetensors",
384
+ "transformer.layers.8.attention.out_proj.weight": "model.safetensors",
385
+ "transformer.layers.8.attention.qkv_proj.biases": "model.safetensors",
386
+ "transformer.layers.8.attention.qkv_proj.scales": "model.safetensors",
387
+ "transformer.layers.8.attention.qkv_proj.weight": "model.safetensors",
388
+ "transformer.layers.8.feed_forward.down_proj.biases": "model.safetensors",
389
+ "transformer.layers.8.feed_forward.down_proj.scales": "model.safetensors",
390
+ "transformer.layers.8.feed_forward.down_proj.weight": "model.safetensors",
391
+ "transformer.layers.8.feed_forward.gate_proj.biases": "model.safetensors",
392
+ "transformer.layers.8.feed_forward.gate_proj.scales": "model.safetensors",
393
+ "transformer.layers.8.feed_forward.gate_proj.weight": "model.safetensors",
394
+ "transformer.layers.8.feed_forward.norm.weight": "model.safetensors",
395
+ "transformer.layers.8.feed_forward.up_proj.biases": "model.safetensors",
396
+ "transformer.layers.8.feed_forward.up_proj.scales": "model.safetensors",
397
+ "transformer.layers.8.feed_forward.up_proj.weight": "model.safetensors",
398
+ "transformer.layers.9.attention.norm.weight": "model.safetensors",
399
+ "transformer.layers.9.attention.out_proj.biases": "model.safetensors",
400
+ "transformer.layers.9.attention.out_proj.scales": "model.safetensors",
401
+ "transformer.layers.9.attention.out_proj.weight": "model.safetensors",
402
+ "transformer.layers.9.attention.qkv_proj.biases": "model.safetensors",
403
+ "transformer.layers.9.attention.qkv_proj.scales": "model.safetensors",
404
+ "transformer.layers.9.attention.qkv_proj.weight": "model.safetensors",
405
+ "transformer.layers.9.feed_forward.down_proj.biases": "model.safetensors",
406
+ "transformer.layers.9.feed_forward.down_proj.scales": "model.safetensors",
407
+ "transformer.layers.9.feed_forward.down_proj.weight": "model.safetensors",
408
+ "transformer.layers.9.feed_forward.gate_proj.biases": "model.safetensors",
409
+ "transformer.layers.9.feed_forward.gate_proj.scales": "model.safetensors",
410
+ "transformer.layers.9.feed_forward.gate_proj.weight": "model.safetensors",
411
+ "transformer.layers.9.feed_forward.norm.weight": "model.safetensors",
412
+ "transformer.layers.9.feed_forward.up_proj.biases": "model.safetensors",
413
+ "transformer.layers.9.feed_forward.up_proj.scales": "model.safetensors",
414
+ "transformer.layers.9.feed_forward.up_proj.weight": "model.safetensors",
415
+ "transformer.norm.weight": "model.safetensors",
416
+ "transformer.tok_embeddings.biases": "model.safetensors",
417
+ "transformer.tok_embeddings.scales": "model.safetensors",
418
+ "transformer.tok_embeddings.weight": "model.safetensors"
419
+ }
420
+ }
special_tokens_map.json ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|assistant|>",
4
+ "<|user|>"
5
+ ],
6
+ "bos_token": {
7
+ "content": "<|startoftext|>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false
12
+ },
13
+ "eos_token": {
14
+ "content": "<|endoftext|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false
19
+ },
20
+ "pad_token": {
21
+ "content": "<|pad|>",
22
+ "lstrip": false,
23
+ "normalized": false,
24
+ "rstrip": false,
25
+ "single_word": false
26
+ },
27
+ "unk_token": {
28
+ "content": "<|endoftext|>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false
33
+ }
34
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,59 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": false,
3
+ "add_prefix_space": true,
4
+ "added_tokens_decoder": {
5
+ "32763": {
6
+ "content": "<|pad|>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "32764": {
14
+ "content": "<|endoftext|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "32765": {
22
+ "content": "<|assistant|>",
23
+ "lstrip": false,
24
+ "normalized": false,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ },
29
+ "32766": {
30
+ "content": "<|user|>",
31
+ "lstrip": false,
32
+ "normalized": false,
33
+ "rstrip": false,
34
+ "single_word": false,
35
+ "special": true
36
+ },
37
+ "32767": {
38
+ "content": "<|startoftext|>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false,
43
+ "special": true
44
+ }
45
+ },
46
+ "additional_special_tokens": [
47
+ "<|assistant|>",
48
+ "<|user|>"
49
+ ],
50
+ "bos_token": "<|startoftext|>",
51
+ "clean_up_tokenization_spaces": false,
52
+ "eos_token": "<|endoftext|>",
53
+ "errors": "replace",
54
+ "extra_special_tokens": {},
55
+ "model_max_length": 512,
56
+ "pad_token": "<|pad|>",
57
+ "tokenizer_class": "GPT2Tokenizer",
58
+ "unk_token": "<|endoftext|>"
59
+ }
vocab.json ADDED
The diff for this file is too large to render. See raw diff