File size: 11,891 Bytes
e2d3fc9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
llama_model_loader: loaded meta data with 36 key-value pairs and 323 tensors from xLAM-8x7b-r-IMat-GGUF/xLAM-8x7b-r.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              = llama
llama_model_loader: - kv   1:                               general.type str              = model
llama_model_loader: - kv   2:                               general.name str              = xLAM 8x7b R
llama_model_loader: - kv   3:                           general.finetune str              = r
llama_model_loader: - kv   4:                           general.basename str              = xLAM
llama_model_loader: - kv   5:                         general.size_label str              = 8x7B
llama_model_loader: - kv   6:                            general.license str              = cc-by-nc-4.0
llama_model_loader: - kv   7:                               general.tags arr[str,6]       = ["function-calling", "LLM Agent", "to...
llama_model_loader: - kv   8:                          general.languages arr[str,1]       = ["en"]
llama_model_loader: - kv   9:                           general.datasets arr[str,1]       = ["Salesforce/xlam-function-calling-60k"]
llama_model_loader: - kv  10:                          llama.block_count u32              = 32
llama_model_loader: - kv  11:                       llama.context_length u32              = 32768
llama_model_loader: - kv  12:                     llama.embedding_length u32              = 4096
llama_model_loader: - kv  13:                  llama.feed_forward_length u32              = 14336
llama_model_loader: - kv  14:                 llama.attention.head_count u32              = 32
llama_model_loader: - kv  15:              llama.attention.head_count_kv u32              = 8
llama_model_loader: - kv  16:                       llama.rope.freq_base f32              = 1000000.000000
llama_model_loader: - kv  17:     llama.attention.layer_norm_rms_epsilon f32              = 0.000010
llama_model_loader: - kv  18:                         llama.expert_count u32              = 8
llama_model_loader: - kv  19:                    llama.expert_used_count u32              = 2
llama_model_loader: - kv  20:                          general.file_type u32              = 7
llama_model_loader: - kv  21:                           llama.vocab_size u32              = 32000
llama_model_loader: - kv  22:                 llama.rope.dimension_count u32              = 128
llama_model_loader: - kv  23:            tokenizer.ggml.add_space_prefix bool             = false
llama_model_loader: - kv  24:                       tokenizer.ggml.model str              = llama
llama_model_loader: - kv  25:                         tokenizer.ggml.pre str              = default
llama_model_loader: - kv  26:                      tokenizer.ggml.tokens arr[str,32000]   = ["<unk>", "<s>", "</s>", "<0x00>", "<...
llama_model_loader: - kv  27:                      tokenizer.ggml.scores arr[f32,32000]   = [-1000.000000, -1000.000000, -1000.00...
llama_model_loader: - kv  28:                  tokenizer.ggml.token_type arr[i32,32000]   = [3, 3, 3, 6, 6, 6, 6, 6, 6, 6, 6, 6, ...
llama_model_loader: - kv  29:                tokenizer.ggml.bos_token_id u32              = 1
llama_model_loader: - kv  30:                tokenizer.ggml.eos_token_id u32              = 2
llama_model_loader: - kv  31:            tokenizer.ggml.unknown_token_id u32              = 0
llama_model_loader: - kv  32:               tokenizer.ggml.add_bos_token bool             = false
llama_model_loader: - kv  33:               tokenizer.ggml.add_eos_token bool             = false
llama_model_loader: - kv  34:                    tokenizer.chat_template str              = {%- if messages[0]['role'] == 'system...
llama_model_loader: - kv  35:               general.quantization_version u32              = 2
llama_model_loader: - type  f32:   97 tensors
llama_model_loader: - type q8_0:  226 tensors
llm_load_vocab: special tokens cache size = 3
llm_load_vocab: token to piece cache size = 0.1637 MB
llm_load_print_meta: format           = GGUF V3 (latest)
llm_load_print_meta: arch             = llama
llm_load_print_meta: vocab type       = SPM
llm_load_print_meta: n_vocab          = 32000
llm_load_print_meta: n_merges         = 0
llm_load_print_meta: vocab_only       = 0
llm_load_print_meta: n_ctx_train      = 32768
llm_load_print_meta: n_embd           = 4096
llm_load_print_meta: n_layer          = 32
llm_load_print_meta: n_head           = 32
llm_load_print_meta: n_head_kv        = 8
llm_load_print_meta: n_rot            = 128
llm_load_print_meta: n_swa            = 0
llm_load_print_meta: n_embd_head_k    = 128
llm_load_print_meta: n_embd_head_v    = 128
llm_load_print_meta: n_gqa            = 4
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-05
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             = 14336
llm_load_print_meta: n_expert         = 8
llm_load_print_meta: n_expert_used    = 2
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     = 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: ssm_dt_b_c_rms   = 0
llm_load_print_meta: model type       = 8x7B
llm_load_print_meta: model ftype      = Q8_0
llm_load_print_meta: model params     = 46.70 B
llm_load_print_meta: model size       = 46.22 GiB (8.50 BPW) 
llm_load_print_meta: general.name     = xLAM 8x7b R
llm_load_print_meta: BOS token        = 1 '<s>'
llm_load_print_meta: EOS token        = 2 '</s>'
llm_load_print_meta: UNK token        = 0 '<unk>'
llm_load_print_meta: LF token         = 13 '<0x0A>'
llm_load_print_meta: max token length = 48
ggml_cuda_init: GGML_CUDA_FORCE_MMQ:    no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
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 =    0.29 MiB
llm_load_tensors: offloading 14 repeating layers to GPU
llm_load_tensors: offloaded 14/33 layers to GPU
llm_load_tensors:        CPU buffer size = 47326.64 MiB
llm_load_tensors:      CUDA0 buffer size = 20589.19 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 =    36.00 MiB
llama_kv_cache_init:      CUDA0 KV buffer size =    28.00 MiB
llama_new_context_with_model: KV self size  =   64.00 MiB, K (f16):   32.00 MiB, V (f16):   32.00 MiB
llama_new_context_with_model:  CUDA_Host  output buffer size =     0.12 MiB
llama_new_context_with_model:      CUDA0 compute buffer size =   621.00 MiB
llama_new_context_with_model:  CUDA_Host compute buffer size =     9.01 MiB
llama_new_context_with_model: graph nodes  = 1510
llama_new_context_with_model: graph splits = 220

system_info: n_threads = 25 (n_threads_batch = 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 94.132 ms
compute_imatrix: computing over 148 chunks with batch_size 512
compute_imatrix: 2.73 seconds per pass - ETA 6.72 minutes
[1]3.2091,[2]2.5764,[3]2.6670,[4]2.7400,[5]3.0794,[6]3.0442,[7]2.7896,[8]3.1600,[9]3.2348,
save_imatrix: stored collected data after 10 chunks in xLAM-8x7b-r-IMat-GGUF/imatrix.dat
[10]3.6092,[11]3.7451,[12]3.5304,[13]3.7238,[14]3.9647,[15]4.2845,[16]4.4296,[17]4.5914,[18]4.6927,[19]4.7570,
save_imatrix: stored collected data after 20 chunks in xLAM-8x7b-r-IMat-GGUF/imatrix.dat
[20]4.8856,[21]4.8333,[22]4.6749,[23]4.8031,[24]4.7572,[25]4.7618,[26]4.6138,[27]4.7839,[28]4.6925,[29]4.7802,
save_imatrix: stored collected data after 30 chunks in xLAM-8x7b-r-IMat-GGUF/imatrix.dat
[30]4.6565,[31]4.5694,[32]4.4464,[33]4.3287,[34]4.3669,[35]4.3448,[36]4.2097,[37]4.1172,[38]4.0459,[39]3.9975,
save_imatrix: stored collected data after 40 chunks in xLAM-8x7b-r-IMat-GGUF/imatrix.dat
[40]3.9660,[41]3.9756,[42]3.9421,[43]3.9288,[44]3.8875,[45]3.8719,[46]3.8937,[47]3.8857,[48]3.9613,[49]3.9866,
save_imatrix: stored collected data after 50 chunks in xLAM-8x7b-r-IMat-GGUF/imatrix.dat
[50]3.9215,[51]3.8327,[52]3.8007,[53]3.8036,[54]3.8267,[55]3.8089,[56]3.8003,[57]3.8687,[58]3.9379,[59]3.9762,
save_imatrix: stored collected data after 60 chunks in xLAM-8x7b-r-IMat-GGUF/imatrix.dat
[60]3.9515,[61]3.9584,[62]3.9914,[63]4.0286,[64]4.0969,[65]4.1267,[66]4.1606,[67]4.1925,[68]4.2290,[69]4.2514,
save_imatrix: stored collected data after 70 chunks in xLAM-8x7b-r-IMat-GGUF/imatrix.dat
[70]4.2529,[71]4.2157,[72]4.1911,[73]4.1892,[74]4.2009,[75]4.2373,[76]4.2392,[77]4.2655,[78]4.2845,[79]4.2767,
save_imatrix: stored collected data after 80 chunks in xLAM-8x7b-r-IMat-GGUF/imatrix.dat
[80]4.2793,[81]4.2705,[82]4.2883,[83]4.3023,[84]4.3108,[85]4.3300,[86]4.3293,[87]4.3292,[88]4.3255,[89]4.3388,
save_imatrix: stored collected data after 90 chunks in xLAM-8x7b-r-IMat-GGUF/imatrix.dat
[90]4.3318,[91]4.3218,[92]4.3130,[93]4.3091,[94]4.3297,[95]4.3524,[96]4.3495,[97]4.3517,[98]4.3467,[99]4.3760,
save_imatrix: stored collected data after 100 chunks in xLAM-8x7b-r-IMat-GGUF/imatrix.dat
[100]4.3427,[101]4.3451,[102]4.3358,[103]4.3504,[104]4.3666,[105]4.3633,[106]4.3460,[107]4.3235,[108]4.3017,[109]4.2777,
save_imatrix: stored collected data after 110 chunks in xLAM-8x7b-r-IMat-GGUF/imatrix.dat
[110]4.2546,[111]4.2344,[112]4.2140,[113]4.1946,[114]4.1741,[115]4.1523,[116]4.1599,[117]4.1781,[118]4.2199,[119]4.2643,
save_imatrix: stored collected data after 120 chunks in xLAM-8x7b-r-IMat-GGUF/imatrix.dat
[120]4.3022,[121]4.3614,[122]4.4127,[123]4.4204,[124]4.4323,[125]4.4065,[126]4.4017,[127]4.3960,[128]4.3967,[129]4.3648,
save_imatrix: stored collected data after 130 chunks in xLAM-8x7b-r-IMat-GGUF/imatrix.dat
[130]4.3317,[131]4.3546,[132]4.3780,[133]4.3820,[134]4.3796,[135]4.3935,[136]4.4164,[137]4.4227,[138]4.4354,[139]4.4547,
save_imatrix: stored collected data after 140 chunks in xLAM-8x7b-r-IMat-GGUF/imatrix.dat
[140]4.4655,[141]4.4630,[142]4.4731,[143]4.4571,[144]4.4257,[145]4.4329,[146]4.4292,[147]4.4230,[148]4.4107,
save_imatrix: stored collected data after 148 chunks in xLAM-8x7b-r-IMat-GGUF/imatrix.dat

llama_print_timings:        load time =   18115.96 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 =  377344.94 ms / 75776 tokens (    4.98 ms per token,   200.81 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 =  393295.61 ms / 75777 tokens

Final estimate: PPL = 4.4107 +/- 0.04800