File size: 10,852 Bytes
d6d9696 |
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 |
build: 3785 (64c6af31) with cc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 for x86_64-linux-gnu
llama_model_loader: loaded meta data with 38 key-value pairs and 507 tensors from Mistral-Small-Instruct-2409-IMat-GGUF/Mistral-Small-Instruct-2409.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 = Mistral Small Instruct 2409
llama_model_loader: - kv 3: general.version str = 2409
llama_model_loader: - kv 4: general.finetune str = Instruct
llama_model_loader: - kv 5: general.basename str = Mistral
llama_model_loader: - kv 6: general.size_label str = Small
llama_model_loader: - kv 7: general.license str = other
llama_model_loader: - kv 8: general.license.name str = mrl
llama_model_loader: - kv 9: general.license.link str = https://mistral.ai/licenses/MRL-0.1.md
llama_model_loader: - kv 10: general.languages arr[str,10] = ["en", "fr", "de", "es", "it", "pt", ...
llama_model_loader: - kv 11: llama.block_count u32 = 56
llama_model_loader: - kv 12: llama.context_length u32 = 131072
llama_model_loader: - kv 13: llama.embedding_length u32 = 6144
llama_model_loader: - kv 14: llama.feed_forward_length u32 = 16384
llama_model_loader: - kv 15: llama.attention.head_count u32 = 48
llama_model_loader: - kv 16: llama.attention.head_count_kv u32 = 8
llama_model_loader: - kv 17: llama.rope.freq_base f32 = 1000000.000000
llama_model_loader: - kv 18: llama.attention.layer_norm_rms_epsilon f32 = 0.000010
llama_model_loader: - kv 19: llama.attention.key_length u32 = 128
llama_model_loader: - kv 20: llama.attention.value_length u32 = 128
llama_model_loader: - kv 21: general.file_type u32 = 7
llama_model_loader: - kv 22: llama.vocab_size u32 = 32768
llama_model_loader: - kv 23: llama.rope.dimension_count u32 = 128
llama_model_loader: - kv 24: tokenizer.ggml.add_space_prefix bool = false
llama_model_loader: - kv 25: tokenizer.ggml.model str = llama
llama_model_loader: - kv 26: tokenizer.ggml.pre str = default
llama_model_loader: - kv 27: tokenizer.ggml.tokens arr[str,32768] = ["<unk>", "<s>", "</s>", "[INST]", "[...
llama_model_loader: - kv 28: tokenizer.ggml.scores arr[f32,32768] = [-1000.000000, -1000.000000, -1000.00...
llama_model_loader: - kv 29: tokenizer.ggml.token_type arr[i32,32768] = [3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, ...
llama_model_loader: - kv 30: tokenizer.ggml.bos_token_id u32 = 1
llama_model_loader: - kv 31: tokenizer.ggml.eos_token_id u32 = 2
llama_model_loader: - kv 32: tokenizer.ggml.unknown_token_id u32 = 0
llama_model_loader: - kv 33: tokenizer.ggml.padding_token_id u32 = 2
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 messages[0]["role"] == "system...
llama_model_loader: - kv 37: general.quantization_version u32 = 2
llama_model_loader: - type f32: 113 tensors
llama_model_loader: - type q8_0: 394 tensors
llm_load_vocab: special tokens cache size = 771
llm_load_vocab: token to piece cache size = 0.1732 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 = 32768
llm_load_print_meta: n_merges = 0
llm_load_print_meta: vocab_only = 0
llm_load_print_meta: n_ctx_train = 131072
llm_load_print_meta: n_embd = 6144
llm_load_print_meta: n_layer = 56
llm_load_print_meta: n_head = 48
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 = 6
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 = 16384
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 = 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 = 131072
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 = ?B
llm_load_print_meta: model ftype = Q8_0
llm_load_print_meta: model params = 22.25 B
llm_load_print_meta: model size = 22.02 GiB (8.50 BPW)
llm_load_print_meta: general.name = Mistral Small Instruct 2409
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: PAD token = 2 '</s>'
llm_load_print_meta: LF token = 781 '<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 = 0.47 MiB
llm_load_tensors: offloading 33 repeating layers to GPU
llm_load_tensors: offloaded 33/57 layers to GPU
llm_load_tensors: CPU buffer size = 22544.65 MiB
llm_load_tensors: CUDA0 buffer size = 13044.80 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 = 46.00 MiB
llama_kv_cache_init: CUDA0 KV buffer size = 66.00 MiB
llama_new_context_with_model: KV self size = 112.00 MiB, K (f16): 56.00 MiB, V (f16): 56.00 MiB
llama_new_context_with_model: CUDA_Host output buffer size = 0.12 MiB
llama_new_context_with_model: CUDA0 compute buffer size = 280.00 MiB
llama_new_context_with_model: CUDA_Host compute buffer size = 13.01 MiB
llama_new_context_with_model: graph nodes = 1798
llama_new_context_with_model: graph splits = 257
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 | RISCV_VECT = 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 94.793 ms
compute_imatrix: computing over 148 chunks with batch_size 512
compute_imatrix: 2.04 seconds per pass - ETA 5.02 minutes
[1]3.2578,[2]2.5472,[3]2.6136,[4]2.6874,[5]3.0332,[6]3.0026,[7]2.7358,[8]3.1186,[9]3.2092,[10]3.5380,[11]3.6618,[12]3.4629,[13]3.6335,[14]3.8547,[15]4.1476,[16]4.2618,[17]4.4422,[18]4.5426,[19]4.6160,[20]4.7419,[21]4.6982,[22]4.5494,[23]4.6684,[24]4.6233,[25]4.6366,[26]4.5111,[27]4.6653,[28]4.6147,[29]4.6997,[30]4.5839,[31]4.4966,[32]4.4114,[33]4.4281,[34]4.4607,[35]4.4212,[36]4.2823,[37]4.1864,[38]4.1108,[39]4.0595,[40]4.0172,[41]4.0269,[42]4.0019,[43]3.9943,[44]3.9540,[45]3.9290,[46]3.9498,[47]3.9423,[48]4.0207,[49]4.0451,[50]4.0061,[51]3.9246,[52]3.9374,[53]3.9481,[54]3.9571,[55]3.9401,[56]3.9288,[57]3.9968,[58]4.0671,[59]4.0987,[60]4.0662,[61]4.0831,[62]4.1117,[63]4.1505,[64]4.2216,[65]4.2418,[66]4.2755,[67]4.3050,[68]4.3357,[69]4.3516,[70]4.3645,[71]4.3298,[72]4.3032,[73]4.2990,[74]4.3122,[75]4.3464,[76]4.3419,[77]4.3671,[78]4.3831,[79]4.3728,[80]4.3743,[81]4.3641,[82]4.3770,[83]4.3886,[84]4.3924,[85]4.4106,[86]4.4055,[87]4.4011,[88]4.3942,[89]4.4035,[90]4.3951,[91]4.3805,[92]4.3712,[93]4.3645,[94]4.3920,[95]4.4133,[96]4.4055,[97]4.4084,[98]4.4046,[99]4.4303,[100]4.3983,[101]4.3979,[102]4.3878,[103]4.4028,[104]4.4163,[105]4.4130,[106]4.3927,[107]4.3670,[108]4.3450,[109]4.3203,[110]4.2957,[111]4.2729,[112]4.2506,[113]4.2273,[114]4.2036,[115]4.1819,[116]4.1885,[117]4.2074,[118]4.2497,[119]4.2920,[120]4.3299,[121]4.3881,[122]4.4360,[123]4.4427,[124]4.4511,[125]4.4282,[126]4.4220,[127]4.4146,[128]4.4161,[129]4.3885,[130]4.3600,[131]4.3846,[132]4.4114,[133]4.4152,[134]4.4138,[135]4.4274,[136]4.4498,[137]4.4582,[138]4.4699,[139]4.4877,[140]4.5007,[141]4.4997,[142]4.5229,[143]4.5358,[144]4.5407,[145]4.5531,[146]4.5524,[147]4.5759,[148]4.5945,
Final estimate: PPL = 4.5945 +/- 0.05201
llama_perf_context_print: load time = 4372.23 ms
llama_perf_context_print: prompt eval time = 283454.45 ms / 75776 tokens ( 3.74 ms per token, 267.33 tokens per second)
llama_perf_context_print: eval time = 0.00 ms / 1 runs ( 0.00 ms per token, inf tokens per second)
llama_perf_context_print: total time = 286343.85 ms / 75777 tokens
|