File size: 9,986 Bytes
8015301
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
llama_model_loader: loaded meta data with 21 key-value pairs and 963 tensors from Qwen2-72B-Instruct-IMat-GGUF/Qwen2-72B-Instruct.Q8_0.gguf.hardlink.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              = qwen2
llama_model_loader: - kv   1:                               general.name str              = Qwen2-72B-Instruct
llama_model_loader: - kv   2:                          qwen2.block_count u32              = 80
llama_model_loader: - kv   3:                       qwen2.context_length u32              = 32768
llama_model_loader: - kv   4:                     qwen2.embedding_length u32              = 8192
llama_model_loader: - kv   5:                  qwen2.feed_forward_length u32              = 29568
llama_model_loader: - kv   6:                 qwen2.attention.head_count u32              = 64
llama_model_loader: - kv   7:              qwen2.attention.head_count_kv u32              = 8
llama_model_loader: - kv   8:                       qwen2.rope.freq_base f32              = 1000000.000000
llama_model_loader: - kv   9:     qwen2.attention.layer_norm_rms_epsilon f32              = 0.000001
llama_model_loader: - kv  10:                          general.file_type u32              = 7
llama_model_loader: - kv  11:                       tokenizer.ggml.model str              = gpt2
llama_model_loader: - kv  12:                         tokenizer.ggml.pre str              = qwen2
llama_model_loader: - kv  13:                      tokenizer.ggml.tokens arr[str,152064]  = ["!", "\"", "#", "$", "%", "&", "'", ...
llama_model_loader: - kv  14:                  tokenizer.ggml.token_type arr[i32,152064]  = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv  15:                      tokenizer.ggml.merges arr[str,151387]  = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",...
llama_model_loader: - kv  16:                tokenizer.ggml.eos_token_id u32              = 151645
llama_model_loader: - kv  17:            tokenizer.ggml.padding_token_id u32              = 151643
llama_model_loader: - kv  18:                tokenizer.ggml.bos_token_id u32              = 151643
llama_model_loader: - kv  19:                    tokenizer.chat_template str              = {% for message in messages %}{% if lo...
llama_model_loader: - kv  20:               general.quantization_version u32              = 2
llama_model_loader: - type  f32:  401 tensors
llama_model_loader: - type q8_0:  562 tensors
llm_load_vocab: special tokens cache size = 421
llm_load_vocab: token to piece cache size = 0.9352 MB
llm_load_print_meta: format           = GGUF V3 (latest)
llm_load_print_meta: arch             = qwen2
llm_load_print_meta: vocab type       = BPE
llm_load_print_meta: n_vocab          = 152064
llm_load_print_meta: n_merges         = 151387
llm_load_print_meta: n_ctx_train      = 32768
llm_load_print_meta: n_embd           = 8192
llm_load_print_meta: n_head           = 64
llm_load_print_meta: n_head_kv        = 8
llm_load_print_meta: n_layer          = 80
llm_load_print_meta: n_rot            = 128
llm_load_print_meta: n_embd_head_k    = 128
llm_load_print_meta: n_embd_head_v    = 128
llm_load_print_meta: n_gqa            = 8
llm_load_print_meta: n_embd_k_gqa     = 1024
llm_load_print_meta: n_embd_v_gqa     = 1024
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             = 29568
llm_load_print_meta: n_expert         = 0
llm_load_print_meta: n_expert_used    = 0
llm_load_print_meta: causal attn      = 1
llm_load_print_meta: pooling type     = 0
llm_load_print_meta: rope type        = 2
llm_load_print_meta: rope scaling     = linear
llm_load_print_meta: freq_base_train  = 1000000.0
llm_load_print_meta: freq_scale_train = 1
llm_load_print_meta: n_ctx_orig_yarn  = 32768
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       = 70B
llm_load_print_meta: model ftype      = Q8_0
llm_load_print_meta: model params     = 72.71 B
llm_load_print_meta: model size       = 71.95 GiB (8.50 BPW) 
llm_load_print_meta: general.name     = Qwen2-72B-Instruct
llm_load_print_meta: BOS token        = 151643 '<|endoftext|>'
llm_load_print_meta: EOS token        = 151645 '<|im_end|>'
llm_load_print_meta: PAD token        = 151643 '<|endoftext|>'
llm_load_print_meta: LF token         = 148848 'ÄĬ'
llm_load_print_meta: EOT token        = 151645 '<|im_end|>'
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 =    0.92 MiB
llm_load_tensors: offloading 20 repeating layers to GPU
llm_load_tensors: offloaded 20/81 layers to GPU
llm_load_tensors:        CPU buffer size = 73677.66 MiB
llm_load_tensors:      CUDA0 buffer size = 17788.29 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  = 1000000.0
llama_new_context_with_model: freq_scale = 1
llama_kv_cache_init:  CUDA_Host KV buffer size =   120.00 MiB
llama_kv_cache_init:      CUDA0 KV buffer size =    40.00 MiB
llama_new_context_with_model: KV self size  =  160.00 MiB, K (f16):   80.00 MiB, V (f16):   80.00 MiB
llama_new_context_with_model:  CUDA_Host  output buffer size =     0.58 MiB
llama_new_context_with_model:      CUDA0 compute buffer size =  1575.25 MiB
llama_new_context_with_model:  CUDA_Host compute buffer size =    17.01 MiB
llama_new_context_with_model: graph nodes  = 2806
llama_new_context_with_model: graph splits = 844

system_info: n_threads = 32 / 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 136.611 ms
compute_imatrix: computing over 128 chunks with batch_size 512
compute_imatrix: 27.81 seconds per pass - ETA 59.32 minutes
[1]4.0667,[2]2.9968,[3]2.7877,[4]3.0396,[5]3.0193,[6]2.7974,[7]2.9432,[8]3.0042,[9]3.3837,
save_imatrix: stored collected data after 10 chunks in Qwen2-72B-Instruct-IMat-GGUF/imatrix.dat
[10]3.3551,[11]3.3012,[12]3.6225,[13]3.9980,[14]4.2226,[15]4.5451,[16]4.8045,[17]4.9530,[18]5.1710,[19]5.0171,
save_imatrix: stored collected data after 20 chunks in Qwen2-72B-Instruct-IMat-GGUF/imatrix.dat
[20]5.1555,[21]5.1434,[22]5.1624,[23]5.1016,[24]5.2596,[25]5.3729,[26]5.2817,[27]5.3150,[28]5.3345,[29]5.4403,
save_imatrix: stored collected data after 30 chunks in Qwen2-72B-Instruct-IMat-GGUF/imatrix.dat
[30]5.3995,[31]5.2560,[32]5.0742,[33]4.9801,[34]4.9146,[35]4.8823,[36]4.8909,[37]4.9080,[38]4.9352,[39]4.9022,
save_imatrix: stored collected data after 40 chunks in Qwen2-72B-Instruct-IMat-GGUF/imatrix.dat
[40]5.0046,[41]5.0428,[42]5.1636,[43]5.0551,[44]5.1610,[45]5.2157,[46]5.2933,[47]5.2076,[48]5.2499,[49]5.3324,
save_imatrix: stored collected data after 50 chunks in Qwen2-72B-Instruct-IMat-GGUF/imatrix.dat
[50]5.3917,[51]5.3295,[52]5.4023,[53]5.4976,[54]5.5736,[55]5.6064,[56]5.6562,[57]5.6968,[58]5.7409,[59]5.7721,
save_imatrix: stored collected data after 60 chunks in Qwen2-72B-Instruct-IMat-GGUF/imatrix.dat
[60]5.7922,[61]5.7742,[62]5.7481,[63]5.7837,[64]5.8379,[65]5.8041,[66]5.8132,[67]5.8367,[68]5.7894,[69]5.7512,
save_imatrix: stored collected data after 70 chunks in Qwen2-72B-Instruct-IMat-GGUF/imatrix.dat
[70]5.7483,[71]5.7360,[72]5.7379,[73]5.7505,[74]5.7101,[75]5.6809,[76]5.6496,[77]5.6416,[78]5.6378,[79]5.6251,
save_imatrix: stored collected data after 80 chunks in Qwen2-72B-Instruct-IMat-GGUF/imatrix.dat
[80]5.5726,[81]5.6016,[82]5.6046,[83]5.5769,[84]5.5958,[85]5.6151,[86]5.5989,[87]5.5806,[88]5.5546,[89]5.5692,
save_imatrix: stored collected data after 90 chunks in Qwen2-72B-Instruct-IMat-GGUF/imatrix.dat
[90]5.5935,[91]5.5874,[92]5.5595,[93]5.5326,[94]5.4937,[95]5.4613,[96]5.4304,[97]5.3971,[98]5.3679,[99]5.3492,
save_imatrix: stored collected data after 100 chunks in Qwen2-72B-Instruct-IMat-GGUF/imatrix.dat
[100]5.3608,[101]5.3831,[102]5.4434,[103]5.5090,[104]5.5541,[105]5.6454,[106]5.6999,[107]5.7232,[108]5.7158,[109]5.7137,
save_imatrix: stored collected data after 110 chunks in Qwen2-72B-Instruct-IMat-GGUF/imatrix.dat
[110]5.7052,[111]5.6696,[112]5.6187,[113]5.5809,[114]5.6217,[115]5.6262,[116]5.6315,[117]5.6514,[118]5.6844,[119]5.6896,
save_imatrix: stored collected data after 120 chunks in Qwen2-72B-Instruct-IMat-GGUF/imatrix.dat
[120]5.6892,[121]5.7048,[122]5.6845,[123]5.7023,[124]5.7156,[125]5.7275,[126]5.7575,[127]5.7766,[128]5.8007,
save_imatrix: stored collected data after 128 chunks in Qwen2-72B-Instruct-IMat-GGUF/imatrix.dat

llama_print_timings:        load time =   74513.74 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 = 1161436.60 ms / 65536 tokens (   17.72 ms per token,    56.43 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 = 1210188.40 ms / 65537 tokens

Final estimate: PPL = 5.8007 +/- 0.07472