mofosyne commited on
Commit
a9a7525
β€’
1 Parent(s): 8f0017d

sync llama.cpp

Browse files
.args CHANGED
@@ -1,2 +1,2 @@
1
  -m
2
- Tinyllama-4.6M-v0.0-F16.gguf
 
1
  -m
2
+ TinyLLama-4.6M-v0.0-F16.gguf
Tinyllama-4.6M-v0.0-F16.dump.md β†’ TinyLLama-4.6M-v0.0-F16.dump.md RENAMED
@@ -1,16 +1,16 @@
1
- # Tinyllama-4.6M-v0.0-F16.gguf - GGUF Internal File Dump
2
 
3
  - Endian: LITTLE endian
4
 
5
  ## Key Value Metadata Store
6
 
7
- There are 39 key-value pairs in this file
8
 
9
  | POS | TYPE | Count | Key | Value |
10
  |----:|:----------|------:|:---------------------------------------|:---------------------------------------------------------------------------------|
11
  | 1 | UINT32 | 1 | GGUF.version | 3 |
12
  | 2 | UINT64 | 1 | GGUF.tensor_count | 75 |
13
- | 3 | UINT64 | 1 | GGUF.kv_count | 36 |
14
  | 4 | STRING | 1 | general.architecture | `llama` |
15
  | 5 | STRING | 1 | general.type | `model` |
16
  | 6 | STRING | 1 | general.name | `TinyLLama` |
@@ -23,30 +23,31 @@ There are 39 key-value pairs in this file
23
  | 13 | STRING | 1 | general.license.name | `Apache License Version 2.0, January 2004` |
24
  | 14 | STRING | 1 | general.license.link | `https://huggingface.co/dataset`...`ob/main/markdown/apache-2.0.md` |
25
  | 15 | STRING | 1 | general.url | `https://huggingface.co/mofosyne/TinyLLama-v0-llamafile` |
26
- | 16 | STRING | 1 | general.source.url | `https://huggingface.co/Maykeye/TinyLLama-v0` |
27
- | 17 | STRING | 1 | general.source.repo_url | `https://huggingface.co/Maykeye/TinyLLama-v0` |
28
- | 18 | [STRING] | 5 | general.tags | [ `text generation`, `transformer`, `llama`, `tiny`, `tiny model` ] |
29
- | 19 | [STRING] | 1 | general.languages | [ `en` ] |
30
- | 20 | [STRING] | 2 | general.datasets | [ `https://hugging`...`-GPT4-train.txt`, `https://hugging`...`-GPT4-valid.txt` ] |
31
- | 21 | UINT32 | 1 | llama.block_count | 8 |
32
- | 22 | UINT32 | 1 | llama.context_length | 2048 |
33
- | 23 | UINT32 | 1 | llama.embedding_length | 64 |
34
- | 24 | UINT32 | 1 | llama.feed_forward_length | 256 |
35
- | 25 | UINT32 | 1 | llama.attention.head_count | 16 |
36
- | 26 | FLOAT32 | 1 | llama.attention.layer_norm_rms_epsilon | 1e-06 |
37
- | 27 | UINT32 | 1 | general.file_type | 1 |
38
- | 28 | UINT32 | 1 | llama.vocab_size | 32000 |
39
- | 29 | UINT32 | 1 | llama.rope.dimension_count | 4 |
40
- | 30 | STRING | 1 | tokenizer.ggml.model | `llama` |
41
- | 31 | STRING | 1 | tokenizer.ggml.pre | `default` |
42
- | 32 | [STRING] | 32000 | tokenizer.ggml.tokens | [ `<unk>`, `<s>`, `</s>`, `<0x00>`, `<0x01>`, ... ] |
43
- | 33 | [FLOAT32] | 32000 | tokenizer.ggml.scores | [ 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, ... ] |
44
- | 34 | [INT32] | 32000 | tokenizer.ggml.token_type | [ 2, 3, 3, 6, 6, 6, 6, ... ] |
45
- | 35 | UINT32 | 1 | tokenizer.ggml.bos_token_id | 1 |
46
- | 36 | UINT32 | 1 | tokenizer.ggml.eos_token_id | 2 |
47
- | 37 | UINT32 | 1 | tokenizer.ggml.unknown_token_id | 0 |
48
- | 38 | UINT32 | 1 | tokenizer.ggml.padding_token_id | 0 |
49
- | 39 | UINT32 | 1 | general.quantization_version | 2 |
 
50
 
51
  ## Tensors Overview ~5M Elements
52
 
@@ -68,81 +69,81 @@ This table contains the offset and data segment relative to start of file
68
 
69
  | T_ID | Tensor Layer Name | Data Offset (B) | Data Size (B) |
70
  |-----:|:-------------------------|-----------------:|-----------------:|
71
- | 0 | output.weight | 0xba880 | 0x3e8000 |
72
- | 1 | token_embd.weight | 0x4a2880 | 0x3e8000 |
73
- | 2 | blk.0.attn_norm.weight | 0x88a880 | 0x100 |
74
- | 3 | blk.0.ffn_down.weight | 0x88a980 | 0x8000 |
75
- | 4 | blk.0.ffn_gate.weight | 0x892980 | 0x8000 |
76
- | 5 | blk.0.ffn_up.weight | 0x89a980 | 0x8000 |
77
- | 6 | blk.0.ffn_norm.weight | 0x8a2980 | 0x100 |
78
- | 7 | blk.0.attn_k.weight | 0x8a2a80 | 0x2000 |
79
- | 8 | blk.0.attn_output.weight | 0x8a4a80 | 0x2000 |
80
- | 9 | blk.0.attn_q.weight | 0x8a6a80 | 0x2000 |
81
- | 10 | blk.0.attn_v.weight | 0x8a8a80 | 0x2000 |
82
- | 11 | blk.1.attn_norm.weight | 0x8aaa80 | 0x100 |
83
- | 12 | blk.1.ffn_down.weight | 0x8aab80 | 0x8000 |
84
- | 13 | blk.1.ffn_gate.weight | 0x8b2b80 | 0x8000 |
85
- | 14 | blk.1.ffn_up.weight | 0x8bab80 | 0x8000 |
86
- | 15 | blk.1.ffn_norm.weight | 0x8c2b80 | 0x100 |
87
- | 16 | blk.1.attn_k.weight | 0x8c2c80 | 0x2000 |
88
- | 17 | blk.1.attn_output.weight | 0x8c4c80 | 0x2000 |
89
- | 18 | blk.1.attn_q.weight | 0x8c6c80 | 0x2000 |
90
- | 19 | blk.1.attn_v.weight | 0x8c8c80 | 0x2000 |
91
- | 20 | blk.2.attn_norm.weight | 0x8cac80 | 0x100 |
92
- | 21 | blk.2.ffn_down.weight | 0x8cad80 | 0x8000 |
93
- | 22 | blk.2.ffn_gate.weight | 0x8d2d80 | 0x8000 |
94
- | 23 | blk.2.ffn_up.weight | 0x8dad80 | 0x8000 |
95
- | 24 | blk.2.ffn_norm.weight | 0x8e2d80 | 0x100 |
96
- | 25 | blk.2.attn_k.weight | 0x8e2e80 | 0x2000 |
97
- | 26 | blk.2.attn_output.weight | 0x8e4e80 | 0x2000 |
98
- | 27 | blk.2.attn_q.weight | 0x8e6e80 | 0x2000 |
99
- | 28 | blk.2.attn_v.weight | 0x8e8e80 | 0x2000 |
100
- | 29 | blk.3.attn_norm.weight | 0x8eae80 | 0x100 |
101
- | 30 | blk.3.ffn_down.weight | 0x8eaf80 | 0x8000 |
102
- | 31 | blk.3.ffn_gate.weight | 0x8f2f80 | 0x8000 |
103
- | 32 | blk.3.ffn_up.weight | 0x8faf80 | 0x8000 |
104
- | 33 | blk.3.ffn_norm.weight | 0x902f80 | 0x100 |
105
- | 34 | blk.3.attn_k.weight | 0x903080 | 0x2000 |
106
- | 35 | blk.3.attn_output.weight | 0x905080 | 0x2000 |
107
- | 36 | blk.3.attn_q.weight | 0x907080 | 0x2000 |
108
- | 37 | blk.3.attn_v.weight | 0x909080 | 0x2000 |
109
- | 38 | blk.4.attn_norm.weight | 0x90b080 | 0x100 |
110
- | 39 | blk.4.ffn_down.weight | 0x90b180 | 0x8000 |
111
- | 40 | blk.4.ffn_gate.weight | 0x913180 | 0x8000 |
112
- | 41 | blk.4.ffn_up.weight | 0x91b180 | 0x8000 |
113
- | 42 | blk.4.ffn_norm.weight | 0x923180 | 0x100 |
114
- | 43 | blk.4.attn_k.weight | 0x923280 | 0x2000 |
115
- | 44 | blk.4.attn_output.weight | 0x925280 | 0x2000 |
116
- | 45 | blk.4.attn_q.weight | 0x927280 | 0x2000 |
117
- | 46 | blk.4.attn_v.weight | 0x929280 | 0x2000 |
118
- | 47 | blk.5.attn_norm.weight | 0x92b280 | 0x100 |
119
- | 48 | blk.5.ffn_down.weight | 0x92b380 | 0x8000 |
120
- | 49 | blk.5.ffn_gate.weight | 0x933380 | 0x8000 |
121
- | 50 | blk.5.ffn_up.weight | 0x93b380 | 0x8000 |
122
- | 51 | blk.5.ffn_norm.weight | 0x943380 | 0x100 |
123
- | 52 | blk.5.attn_k.weight | 0x943480 | 0x2000 |
124
- | 53 | blk.5.attn_output.weight | 0x945480 | 0x2000 |
125
- | 54 | blk.5.attn_q.weight | 0x947480 | 0x2000 |
126
- | 55 | blk.5.attn_v.weight | 0x949480 | 0x2000 |
127
- | 56 | blk.6.attn_norm.weight | 0x94b480 | 0x100 |
128
- | 57 | blk.6.ffn_down.weight | 0x94b580 | 0x8000 |
129
- | 58 | blk.6.ffn_gate.weight | 0x953580 | 0x8000 |
130
- | 59 | blk.6.ffn_up.weight | 0x95b580 | 0x8000 |
131
- | 60 | blk.6.ffn_norm.weight | 0x963580 | 0x100 |
132
- | 61 | blk.6.attn_k.weight | 0x963680 | 0x2000 |
133
- | 62 | blk.6.attn_output.weight | 0x965680 | 0x2000 |
134
- | 63 | blk.6.attn_q.weight | 0x967680 | 0x2000 |
135
- | 64 | blk.6.attn_v.weight | 0x969680 | 0x2000 |
136
- | 65 | blk.7.attn_norm.weight | 0x96b680 | 0x100 |
137
- | 66 | blk.7.ffn_down.weight | 0x96b780 | 0x8000 |
138
- | 67 | blk.7.ffn_gate.weight | 0x973780 | 0x8000 |
139
- | 68 | blk.7.ffn_up.weight | 0x97b780 | 0x8000 |
140
- | 69 | blk.7.ffn_norm.weight | 0x983780 | 0x100 |
141
- | 70 | blk.7.attn_k.weight | 0x983880 | 0x2000 |
142
- | 71 | blk.7.attn_output.weight | 0x985880 | 0x2000 |
143
- | 72 | blk.7.attn_q.weight | 0x987880 | 0x2000 |
144
- | 73 | blk.7.attn_v.weight | 0x989880 | 0x2000 |
145
- | 74 | output_norm.weight | 0x98b880 | 0x100 |
146
 
147
  ### <a name="base">Base Tensor Group : ~4M Elements</a>
148
 
 
1
+ # TinyLLama-4.6M-v0.0-F16.gguf - GGUF Internal File Dump
2
 
3
  - Endian: LITTLE endian
4
 
5
  ## Key Value Metadata Store
6
 
7
+ There are 40 key-value pairs in this file
8
 
9
  | POS | TYPE | Count | Key | Value |
10
  |----:|:----------|------:|:---------------------------------------|:---------------------------------------------------------------------------------|
11
  | 1 | UINT32 | 1 | GGUF.version | 3 |
12
  | 2 | UINT64 | 1 | GGUF.tensor_count | 75 |
13
+ | 3 | UINT64 | 1 | GGUF.kv_count | 37 |
14
  | 4 | STRING | 1 | general.architecture | `llama` |
15
  | 5 | STRING | 1 | general.type | `model` |
16
  | 6 | STRING | 1 | general.name | `TinyLLama` |
 
23
  | 13 | STRING | 1 | general.license.name | `Apache License Version 2.0, January 2004` |
24
  | 14 | STRING | 1 | general.license.link | `https://huggingface.co/dataset`...`ob/main/markdown/apache-2.0.md` |
25
  | 15 | STRING | 1 | general.url | `https://huggingface.co/mofosyne/TinyLLama-v0-llamafile` |
26
+ | 16 | STRING | 1 | general.repo_url | `https://huggingface.co/mofosyne/TinyLLama-v0-llamafile` |
27
+ | 17 | STRING | 1 | general.source.url | `https://huggingface.co/Maykeye/TinyLLama-v0` |
28
+ | 18 | STRING | 1 | general.source.repo_url | `https://huggingface.co/Maykeye/TinyLLama-v0` |
29
+ | 19 | [STRING] | 5 | general.tags | [ `text generation`, `transformer`, `llama`, `tiny`, `tiny model` ] |
30
+ | 20 | [STRING] | 1 | general.languages | [ `en` ] |
31
+ | 21 | [STRING] | 2 | general.datasets | [ `https://hugging`...`-GPT4-train.txt`, `https://hugging`...`-GPT4-valid.txt` ] |
32
+ | 22 | UINT32 | 1 | llama.block_count | 8 |
33
+ | 23 | UINT32 | 1 | llama.context_length | 2048 |
34
+ | 24 | UINT32 | 1 | llama.embedding_length | 64 |
35
+ | 25 | UINT32 | 1 | llama.feed_forward_length | 256 |
36
+ | 26 | UINT32 | 1 | llama.attention.head_count | 16 |
37
+ | 27 | FLOAT32 | 1 | llama.attention.layer_norm_rms_epsilon | 1e-06 |
38
+ | 28 | UINT32 | 1 | general.file_type | 1 |
39
+ | 29 | UINT32 | 1 | llama.vocab_size | 32000 |
40
+ | 30 | UINT32 | 1 | llama.rope.dimension_count | 4 |
41
+ | 31 | STRING | 1 | tokenizer.ggml.model | `llama` |
42
+ | 32 | STRING | 1 | tokenizer.ggml.pre | `default` |
43
+ | 33 | [STRING] | 32000 | tokenizer.ggml.tokens | [ `<unk>`, `<s>`, `</s>`, `<0x00>`, `<0x01>`, ... ] |
44
+ | 34 | [FLOAT32] | 32000 | tokenizer.ggml.scores | [ 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, ... ] |
45
+ | 35 | [INT32] | 32000 | tokenizer.ggml.token_type | [ 2, 3, 3, 6, 6, 6, 6, ... ] |
46
+ | 36 | UINT32 | 1 | tokenizer.ggml.bos_token_id | 1 |
47
+ | 37 | UINT32 | 1 | tokenizer.ggml.eos_token_id | 2 |
48
+ | 38 | UINT32 | 1 | tokenizer.ggml.unknown_token_id | 0 |
49
+ | 39 | UINT32 | 1 | tokenizer.ggml.padding_token_id | 0 |
50
+ | 40 | UINT32 | 1 | general.quantization_version | 2 |
51
 
52
  ## Tensors Overview ~5M Elements
53
 
 
69
 
70
  | T_ID | Tensor Layer Name | Data Offset (B) | Data Size (B) |
71
  |-----:|:-------------------------|-----------------:|-----------------:|
72
+ | 0 | output.weight | 0xba8e0 | 0x3e8000 |
73
+ | 1 | token_embd.weight | 0x4a28e0 | 0x3e8000 |
74
+ | 2 | blk.0.attn_norm.weight | 0x88a8e0 | 0x100 |
75
+ | 3 | blk.0.ffn_down.weight | 0x88a9e0 | 0x8000 |
76
+ | 4 | blk.0.ffn_gate.weight | 0x8929e0 | 0x8000 |
77
+ | 5 | blk.0.ffn_up.weight | 0x89a9e0 | 0x8000 |
78
+ | 6 | blk.0.ffn_norm.weight | 0x8a29e0 | 0x100 |
79
+ | 7 | blk.0.attn_k.weight | 0x8a2ae0 | 0x2000 |
80
+ | 8 | blk.0.attn_output.weight | 0x8a4ae0 | 0x2000 |
81
+ | 9 | blk.0.attn_q.weight | 0x8a6ae0 | 0x2000 |
82
+ | 10 | blk.0.attn_v.weight | 0x8a8ae0 | 0x2000 |
83
+ | 11 | blk.1.attn_norm.weight | 0x8aaae0 | 0x100 |
84
+ | 12 | blk.1.ffn_down.weight | 0x8aabe0 | 0x8000 |
85
+ | 13 | blk.1.ffn_gate.weight | 0x8b2be0 | 0x8000 |
86
+ | 14 | blk.1.ffn_up.weight | 0x8babe0 | 0x8000 |
87
+ | 15 | blk.1.ffn_norm.weight | 0x8c2be0 | 0x100 |
88
+ | 16 | blk.1.attn_k.weight | 0x8c2ce0 | 0x2000 |
89
+ | 17 | blk.1.attn_output.weight | 0x8c4ce0 | 0x2000 |
90
+ | 18 | blk.1.attn_q.weight | 0x8c6ce0 | 0x2000 |
91
+ | 19 | blk.1.attn_v.weight | 0x8c8ce0 | 0x2000 |
92
+ | 20 | blk.2.attn_norm.weight | 0x8cace0 | 0x100 |
93
+ | 21 | blk.2.ffn_down.weight | 0x8cade0 | 0x8000 |
94
+ | 22 | blk.2.ffn_gate.weight | 0x8d2de0 | 0x8000 |
95
+ | 23 | blk.2.ffn_up.weight | 0x8dade0 | 0x8000 |
96
+ | 24 | blk.2.ffn_norm.weight | 0x8e2de0 | 0x100 |
97
+ | 25 | blk.2.attn_k.weight | 0x8e2ee0 | 0x2000 |
98
+ | 26 | blk.2.attn_output.weight | 0x8e4ee0 | 0x2000 |
99
+ | 27 | blk.2.attn_q.weight | 0x8e6ee0 | 0x2000 |
100
+ | 28 | blk.2.attn_v.weight | 0x8e8ee0 | 0x2000 |
101
+ | 29 | blk.3.attn_norm.weight | 0x8eaee0 | 0x100 |
102
+ | 30 | blk.3.ffn_down.weight | 0x8eafe0 | 0x8000 |
103
+ | 31 | blk.3.ffn_gate.weight | 0x8f2fe0 | 0x8000 |
104
+ | 32 | blk.3.ffn_up.weight | 0x8fafe0 | 0x8000 |
105
+ | 33 | blk.3.ffn_norm.weight | 0x902fe0 | 0x100 |
106
+ | 34 | blk.3.attn_k.weight | 0x9030e0 | 0x2000 |
107
+ | 35 | blk.3.attn_output.weight | 0x9050e0 | 0x2000 |
108
+ | 36 | blk.3.attn_q.weight | 0x9070e0 | 0x2000 |
109
+ | 37 | blk.3.attn_v.weight | 0x9090e0 | 0x2000 |
110
+ | 38 | blk.4.attn_norm.weight | 0x90b0e0 | 0x100 |
111
+ | 39 | blk.4.ffn_down.weight | 0x90b1e0 | 0x8000 |
112
+ | 40 | blk.4.ffn_gate.weight | 0x9131e0 | 0x8000 |
113
+ | 41 | blk.4.ffn_up.weight | 0x91b1e0 | 0x8000 |
114
+ | 42 | blk.4.ffn_norm.weight | 0x9231e0 | 0x100 |
115
+ | 43 | blk.4.attn_k.weight | 0x9232e0 | 0x2000 |
116
+ | 44 | blk.4.attn_output.weight | 0x9252e0 | 0x2000 |
117
+ | 45 | blk.4.attn_q.weight | 0x9272e0 | 0x2000 |
118
+ | 46 | blk.4.attn_v.weight | 0x9292e0 | 0x2000 |
119
+ | 47 | blk.5.attn_norm.weight | 0x92b2e0 | 0x100 |
120
+ | 48 | blk.5.ffn_down.weight | 0x92b3e0 | 0x8000 |
121
+ | 49 | blk.5.ffn_gate.weight | 0x9333e0 | 0x8000 |
122
+ | 50 | blk.5.ffn_up.weight | 0x93b3e0 | 0x8000 |
123
+ | 51 | blk.5.ffn_norm.weight | 0x9433e0 | 0x100 |
124
+ | 52 | blk.5.attn_k.weight | 0x9434e0 | 0x2000 |
125
+ | 53 | blk.5.attn_output.weight | 0x9454e0 | 0x2000 |
126
+ | 54 | blk.5.attn_q.weight | 0x9474e0 | 0x2000 |
127
+ | 55 | blk.5.attn_v.weight | 0x9494e0 | 0x2000 |
128
+ | 56 | blk.6.attn_norm.weight | 0x94b4e0 | 0x100 |
129
+ | 57 | blk.6.ffn_down.weight | 0x94b5e0 | 0x8000 |
130
+ | 58 | blk.6.ffn_gate.weight | 0x9535e0 | 0x8000 |
131
+ | 59 | blk.6.ffn_up.weight | 0x95b5e0 | 0x8000 |
132
+ | 60 | blk.6.ffn_norm.weight | 0x9635e0 | 0x100 |
133
+ | 61 | blk.6.attn_k.weight | 0x9636e0 | 0x2000 |
134
+ | 62 | blk.6.attn_output.weight | 0x9656e0 | 0x2000 |
135
+ | 63 | blk.6.attn_q.weight | 0x9676e0 | 0x2000 |
136
+ | 64 | blk.6.attn_v.weight | 0x9696e0 | 0x2000 |
137
+ | 65 | blk.7.attn_norm.weight | 0x96b6e0 | 0x100 |
138
+ | 66 | blk.7.ffn_down.weight | 0x96b7e0 | 0x8000 |
139
+ | 67 | blk.7.ffn_gate.weight | 0x9737e0 | 0x8000 |
140
+ | 68 | blk.7.ffn_up.weight | 0x97b7e0 | 0x8000 |
141
+ | 69 | blk.7.ffn_norm.weight | 0x9837e0 | 0x100 |
142
+ | 70 | blk.7.attn_k.weight | 0x9838e0 | 0x2000 |
143
+ | 71 | blk.7.attn_output.weight | 0x9858e0 | 0x2000 |
144
+ | 72 | blk.7.attn_q.weight | 0x9878e0 | 0x2000 |
145
+ | 73 | blk.7.attn_v.weight | 0x9898e0 | 0x2000 |
146
+ | 74 | output_norm.weight | 0x98b8e0 | 0x100 |
147
 
148
  ### <a name="base">Base Tensor Group : ~4M Elements</a>
149
 
Tinyllama-4.6M-v0.0-F16.gguf β†’ TinyLLama-4.6M-v0.0-F16.gguf RENAMED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:4b8df674f3486214ade8412f6f0bc42168ad37ee086ebe3b1888832a8d8bca7c
3
- size 10008960
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:85075ffdac973729bcec18cd6a4550e3c672c8dedeb0f65133f3f88aa35e1f2a
3
+ size 10009056
Tinyllama-4.6M-v0.0-F16.llamafile β†’ TinyLLama-4.6M-v0.0-F16.llamafile RENAMED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d583b00bc88fc261078db9c5979ea65d2a4a5be4d7bf5c1662c5678ea9e72c28
3
  size 21509955
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c03831b026a42cc43627c6efcd5b382c60af0abcf0bc6c10666cfbf37f568b9d
3
  size 21509955
llama.cpp CHANGED
@@ -1 +1 @@
1
- Subproject commit c3776cacabce2ee35f172fb72be7a519752125fa
 
1
+ Subproject commit 22f281aa16f44d8f6ec2c180a0685ff27e04e714