File size: 12,362 Bytes
6231ed5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
llama_model_loader: loaded meta data with 38 key-value pairs and 377 tensors from DeepSeek-Coder-V2-Lite-Instruct-IMat-GGUF/DeepSeek-Coder-V2-Lite-Instruct.gguf (version GGUF V3 (latest))
llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
llama_model_loader: - kv   0:                       general.architecture str              = deepseek2
llama_model_loader: - kv   1:                               general.name str              = DeepSeek-Coder-V2-Lite-Instruct
llama_model_loader: - kv   2:                      deepseek2.block_count u32              = 27
llama_model_loader: - kv   3:                   deepseek2.context_length u32              = 163840
llama_model_loader: - kv   4:                 deepseek2.embedding_length u32              = 2048
llama_model_loader: - kv   5:              deepseek2.feed_forward_length u32              = 10944
llama_model_loader: - kv   6:             deepseek2.attention.head_count u32              = 16
llama_model_loader: - kv   7:          deepseek2.attention.head_count_kv u32              = 16
llama_model_loader: - kv   8:                   deepseek2.rope.freq_base f32              = 10000.000000
llama_model_loader: - kv   9: deepseek2.attention.layer_norm_rms_epsilon f32              = 0.000001
llama_model_loader: - kv  10:                deepseek2.expert_used_count u32              = 6
llama_model_loader: - kv  11:                          general.file_type u32              = 0
llama_model_loader: - kv  12:        deepseek2.leading_dense_block_count u32              = 1
llama_model_loader: - kv  13:                       deepseek2.vocab_size u32              = 102400
llama_model_loader: - kv  14:           deepseek2.attention.kv_lora_rank u32              = 512
llama_model_loader: - kv  15:             deepseek2.attention.key_length u32              = 192
llama_model_loader: - kv  16:           deepseek2.attention.value_length u32              = 128
llama_model_loader: - kv  17:       deepseek2.expert_feed_forward_length u32              = 1408
llama_model_loader: - kv  18:                     deepseek2.expert_count u32              = 64
llama_model_loader: - kv  19:              deepseek2.expert_shared_count u32              = 2
llama_model_loader: - kv  20:             deepseek2.expert_weights_scale f32              = 1.000000
llama_model_loader: - kv  21:             deepseek2.rope.dimension_count u32              = 64
llama_model_loader: - kv  22:                deepseek2.rope.scaling.type str              = yarn
llama_model_loader: - kv  23:              deepseek2.rope.scaling.factor f32              = 40.000000
llama_model_loader: - kv  24: deepseek2.rope.scaling.original_context_length u32              = 4096
llama_model_loader: - kv  25: deepseek2.rope.scaling.yarn_log_multiplier f32              = 0.070700
llama_model_loader: - kv  26:                       tokenizer.ggml.model str              = gpt2
llama_model_loader: - kv  27:                         tokenizer.ggml.pre str              = deepseek-llm
llama_model_loader: - kv  28:                      tokenizer.ggml.tokens arr[str,102400]  = ["!", "\"", "#", "$", "%", "&", "'", ...
llama_model_loader: - kv  29:                  tokenizer.ggml.token_type arr[i32,102400]  = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv  30:                      tokenizer.ggml.merges arr[str,99757]   = ["Ġ Ġ", "Ġ t", "Ġ a", "i n", "h e...
llama_model_loader: - kv  31:                tokenizer.ggml.bos_token_id u32              = 100000
llama_model_loader: - kv  32:                tokenizer.ggml.eos_token_id u32              = 100001
llama_model_loader: - kv  33:            tokenizer.ggml.padding_token_id u32              = 100001
llama_model_loader: - kv  34:               tokenizer.ggml.add_bos_token bool             = true
llama_model_loader: - kv  35:               tokenizer.ggml.add_eos_token bool             = false
llama_model_loader: - kv  36:                    tokenizer.chat_template str              = {% if not add_generation_prompt is de...
llama_model_loader: - kv  37:               general.quantization_version u32              = 2
llama_model_loader: - type  f32:  377 tensors
llm_load_vocab: special tokens cache size = 2400
llm_load_vocab: token to piece cache size = 0.6661 MB
llm_load_print_meta: format           = GGUF V3 (latest)
llm_load_print_meta: arch             = deepseek2
llm_load_print_meta: vocab type       = BPE
llm_load_print_meta: n_vocab          = 102400
llm_load_print_meta: n_merges         = 99757
llm_load_print_meta: n_ctx_train      = 163840
llm_load_print_meta: n_embd           = 2048
llm_load_print_meta: n_head           = 16
llm_load_print_meta: n_head_kv        = 16
llm_load_print_meta: n_layer          = 27
llm_load_print_meta: n_rot            = 64
llm_load_print_meta: n_embd_head_k    = 192
llm_load_print_meta: n_embd_head_v    = 128
llm_load_print_meta: n_gqa            = 1
llm_load_print_meta: n_embd_k_gqa     = 3072
llm_load_print_meta: n_embd_v_gqa     = 2048
llm_load_print_meta: f_norm_eps       = 0.0e+00
llm_load_print_meta: f_norm_rms_eps   = 1.0e-06
llm_load_print_meta: f_clamp_kqv      = 0.0e+00
llm_load_print_meta: f_max_alibi_bias = 0.0e+00
llm_load_print_meta: f_logit_scale    = 0.0e+00
llm_load_print_meta: n_ff             = 10944
llm_load_print_meta: n_expert         = 64
llm_load_print_meta: n_expert_used    = 6
llm_load_print_meta: causal attn      = 1
llm_load_print_meta: pooling type     = 0
llm_load_print_meta: rope type        = 0
llm_load_print_meta: rope scaling     = yarn
llm_load_print_meta: freq_base_train  = 10000.0
llm_load_print_meta: freq_scale_train = 0.025
llm_load_print_meta: n_ctx_orig_yarn  = 4096
llm_load_print_meta: rope_finetuned   = unknown
llm_load_print_meta: ssm_d_conv       = 0
llm_load_print_meta: ssm_d_inner      = 0
llm_load_print_meta: ssm_d_state      = 0
llm_load_print_meta: ssm_dt_rank      = 0
llm_load_print_meta: model type       = 16B
llm_load_print_meta: model ftype      = all F32
llm_load_print_meta: model params     = 15.71 B
llm_load_print_meta: model size       = 58.51 GiB (32.00 BPW) 
llm_load_print_meta: general.name     = DeepSeek-Coder-V2-Lite-Instruct
llm_load_print_meta: BOS token        = 100000 '<|begin▁of▁sentence|>'
llm_load_print_meta: EOS token        = 100001 '<|end▁of▁sentence|>'
llm_load_print_meta: PAD token        = 100001 '<|end▁of▁sentence|>'
llm_load_print_meta: LF token         = 126 'Ä'
llm_load_print_meta: n_layer_dense_lead   = 1
llm_load_print_meta: n_lora_q             = 0
llm_load_print_meta: n_lora_kv            = 512
llm_load_print_meta: n_ff_exp             = 1408
llm_load_print_meta: n_expert_shared      = 2
llm_load_print_meta: expert_weights_scale = 1.0
llm_load_print_meta: rope_yarn_log_mul    = 0.0707
ggml_cuda_init: GGML_CUDA_FORCE_MMQ:   no
ggml_cuda_init: CUDA_USE_TENSOR_CORES: yes
ggml_cuda_init: found 1 CUDA devices:
  Device 0: NVIDIA GeForce RTX 4090, compute capability 8.9, VMM: yes
llm_load_tensors: ggml ctx size =    0.35 MiB
llm_load_tensors: offloading 10 repeating layers to GPU
llm_load_tensors: offloaded 10/28 layers to GPU
llm_load_tensors:        CPU buffer size = 37605.31 MiB
llm_load_tensors:      CUDA0 buffer size = 22310.18 MiB
.......................................................................................
llama_new_context_with_model: n_ctx      = 512
llama_new_context_with_model: n_batch    = 512
llama_new_context_with_model: n_ubatch   = 512
llama_new_context_with_model: flash_attn = 0
llama_new_context_with_model: freq_base  = 10000.0
llama_new_context_with_model: freq_scale = 0.025
llama_kv_cache_init:  CUDA_Host KV buffer size =    85.00 MiB
llama_kv_cache_init:      CUDA0 KV buffer size =    50.00 MiB
llama_new_context_with_model: KV self size  =  135.00 MiB, K (f16):   81.00 MiB, V (f16):   54.00 MiB
llama_new_context_with_model:  CUDA_Host  output buffer size =     0.39 MiB
llama_new_context_with_model:      CUDA0 compute buffer size =  1012.00 MiB
llama_new_context_with_model:  CUDA_Host compute buffer size =    11.01 MiB
llama_new_context_with_model: graph nodes  = 1924
llama_new_context_with_model: graph splits = 272

system_info: n_threads = 25 / 32 | AVX = 1 | AVX_VNNI = 0 | AVX2 = 1 | AVX512 = 1 | AVX512_VBMI = 1 | AVX512_VNNI = 1 | AVX512_BF16 = 1 | FMA = 1 | NEON = 0 | SVE = 0 | ARM_FMA = 0 | F16C = 1 | FP16_VA = 0 | WASM_SIMD = 0 | BLAS = 1 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 1 | 
compute_imatrix: tokenizing the input ..
compute_imatrix: tokenization took 215.372 ms
compute_imatrix: computing over 139 chunks with batch_size 512
compute_imatrix: 2.97 seconds per pass - ETA 6.88 minutes
[1]8.3398,[2]5.4215,[3]5.2838,[4]6.0526,[5]5.6694,[6]5.3675,[7]5.7148,[8]5.8289,[9]6.5437,
save_imatrix: stored collected data after 10 chunks in DeepSeek-Coder-V2-Lite-Instruct-IMat-GGUF/imatrix.dat
[10]6.6919,[11]6.1685,[12]6.4804,[13]6.9224,[14]7.2573,[15]7.3708,[16]7.7938,[17]8.0288,[18]8.1908,[19]8.5073,
save_imatrix: stored collected data after 20 chunks in DeepSeek-Coder-V2-Lite-Instruct-IMat-GGUF/imatrix.dat
[20]8.0227,[21]7.9922,[22]7.9696,[23]8.1526,[24]7.9689,[25]8.2193,[26]8.0970,[27]8.2980,[28]8.1099,[29]8.3295,
save_imatrix: stored collected data after 30 chunks in DeepSeek-Coder-V2-Lite-Instruct-IMat-GGUF/imatrix.dat
[30]8.5506,[31]8.5974,[32]8.4411,[33]8.1166,[34]7.6767,[35]7.3223,[36]7.2017,[37]7.0882,[38]7.0476,[39]6.9338,
save_imatrix: stored collected data after 40 chunks in DeepSeek-Coder-V2-Lite-Instruct-IMat-GGUF/imatrix.dat
[40]6.8448,[41]6.6842,[42]6.5776,[43]6.6411,[44]6.7147,[45]6.8307,[46]6.8212,[47]7.0831,[48]7.3037,[49]7.4481,
save_imatrix: stored collected data after 50 chunks in DeepSeek-Coder-V2-Lite-Instruct-IMat-GGUF/imatrix.dat
[50]7.5775,[51]7.6764,[52]7.5752,[53]7.7096,[54]7.8061,[55]7.9151,[56]7.8063,[57]7.7795,[58]7.7842,[59]7.8984,
save_imatrix: stored collected data after 60 chunks in DeepSeek-Coder-V2-Lite-Instruct-IMat-GGUF/imatrix.dat
[60]8.0463,[61]8.1568,[62]8.2185,[63]8.2355,[64]8.2746,[65]8.2711,[66]8.2385,[67]8.2192,[68]8.1754,[69]8.2307,
save_imatrix: stored collected data after 70 chunks in DeepSeek-Coder-V2-Lite-Instruct-IMat-GGUF/imatrix.dat
[70]8.2569,[71]8.2581,[72]8.2878,[73]8.2914,[74]8.2632,[75]8.2405,[76]8.2185,[77]8.2144,[78]8.2302,[79]8.1705,
save_imatrix: stored collected data after 80 chunks in DeepSeek-Coder-V2-Lite-Instruct-IMat-GGUF/imatrix.dat
[80]8.2038,[81]8.1894,[82]8.1776,[83]8.1431,[84]8.1184,[85]8.0975,[86]8.0670,[87]8.0352,[88]8.0625,[89]8.0887,
save_imatrix: stored collected data after 90 chunks in DeepSeek-Coder-V2-Lite-Instruct-IMat-GGUF/imatrix.dat
[90]8.0910,[91]8.1317,[92]8.1702,[93]8.1019,[94]8.0817,[95]8.0496,[96]8.0788,[97]8.0812,[98]8.0650,[99]7.9677,
save_imatrix: stored collected data after 100 chunks in DeepSeek-Coder-V2-Lite-Instruct-IMat-GGUF/imatrix.dat
[100]7.8685,[101]7.7737,[102]7.6755,[103]7.5767,[104]7.4886,[105]7.4023,[106]7.3171,[107]7.2332,[108]7.1950,[109]7.2108,
save_imatrix: stored collected data after 110 chunks in DeepSeek-Coder-V2-Lite-Instruct-IMat-GGUF/imatrix.dat
[110]7.2567,[111]7.3272,[112]7.3958,[113]7.4487,[114]7.5589,[115]7.6035,[116]7.6316,[117]7.6348,[118]7.6765,[119]7.6743,
save_imatrix: stored collected data after 120 chunks in DeepSeek-Coder-V2-Lite-Instruct-IMat-GGUF/imatrix.dat
[120]7.6614,[121]7.5726,[122]7.4906,[123]7.5336,[124]7.5893,[125]7.5868,[126]7.5893,[127]7.5967,[128]7.6244,[129]7.6239,
save_imatrix: stored collected data after 130 chunks in DeepSeek-Coder-V2-Lite-Instruct-IMat-GGUF/imatrix.dat
[130]7.6332,[131]7.6606,[132]7.6577,[133]7.6290,[134]7.6764,[135]7.7304,[136]7.7647,[137]7.8148,[138]7.8820,[139]7.9125,
save_imatrix: stored collected data after 139 chunks in DeepSeek-Coder-V2-Lite-Instruct-IMat-GGUF/imatrix.dat

llama_print_timings:        load time =   17517.65 ms
llama_print_timings:      sample time =       0.00 ms /     1 runs   (    0.00 ms per token,      inf tokens per second)
llama_print_timings: prompt eval time =  385218.87 ms / 71168 tokens (    5.41 ms per token,   184.75 tokens per second)
llama_print_timings:        eval time =       0.00 ms /     1 runs   (    0.00 ms per token,      inf tokens per second)
llama_print_timings:       total time =  400724.75 ms / 71169 tokens

Final estimate: PPL = 7.9125 +/- 0.12345