main: build = 3023 (ee3dff6b) main: built with cc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 for x86_64-linux-gnu main: seed = 1716921775 llama_model_loader: loaded meta data with 21 key-value pairs and 291 tensors from Llama-3-Instruct-8B-SimPO-IMat-GGUF/Llama-3-Instruct-8B-SimPO.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.name str = Llama-3-Instruct-8B-SimPO llama_model_loader: - kv 2: llama.block_count u32 = 32 llama_model_loader: - kv 3: llama.context_length u32 = 8192 llama_model_loader: - kv 4: llama.embedding_length u32 = 4096 llama_model_loader: - kv 5: llama.feed_forward_length u32 = 14336 llama_model_loader: - kv 6: llama.attention.head_count u32 = 32 llama_model_loader: - kv 7: llama.attention.head_count_kv u32 = 8 llama_model_loader: - kv 8: llama.rope.freq_base f32 = 500000.000000 llama_model_loader: - kv 9: llama.attention.layer_norm_rms_epsilon f32 = 0.000010 llama_model_loader: - kv 10: general.file_type u32 = 0 llama_model_loader: - kv 11: llama.vocab_size u32 = 128256 llama_model_loader: - kv 12: llama.rope.dimension_count u32 = 128 llama_model_loader: - kv 13: tokenizer.ggml.model str = gpt2 llama_model_loader: - kv 14: tokenizer.ggml.pre str = llama-bpe llama_model_loader: - kv 15: tokenizer.ggml.tokens arr[str,128256] = ["!", "\"", "#", "$", "%", "&", "'", ... llama_model_loader: - kv 16: tokenizer.ggml.token_type arr[i32,128256] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ... llama_model_loader: - kv 17: tokenizer.ggml.merges arr[str,280147] = ["Ġ Ġ", "Ġ ĠĠĠ", "ĠĠ ĠĠ", "... llama_model_loader: - kv 18: tokenizer.ggml.bos_token_id u32 = 128000 llama_model_loader: - kv 19: tokenizer.ggml.eos_token_id u32 = 128001 llama_model_loader: - kv 20: general.quantization_version u32 = 2 llama_model_loader: - type f32: 291 tensors llm_load_vocab: special tokens definition check successful ( 256/128256 ). llm_load_print_meta: format = GGUF V3 (latest) llm_load_print_meta: arch = llama llm_load_print_meta: vocab type = BPE llm_load_print_meta: n_vocab = 128256 llm_load_print_meta: n_merges = 280147 llm_load_print_meta: n_ctx_train = 8192 llm_load_print_meta: n_embd = 4096 llm_load_print_meta: n_head = 32 llm_load_print_meta: n_head_kv = 8 llm_load_print_meta: n_layer = 32 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 = 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 = 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 = 500000.0 llm_load_print_meta: freq_scale_train = 1 llm_load_print_meta: n_yarn_orig_ctx = 8192 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 = 8B llm_load_print_meta: model ftype = all F32 llm_load_print_meta: model params = 8.03 B llm_load_print_meta: model size = 29.92 GiB (32.00 BPW) llm_load_print_meta: general.name = Llama-3-Instruct-8B-SimPO llm_load_print_meta: BOS token = 128000 '<|begin_of_text|>' llm_load_print_meta: EOS token = 128001 '<|end_of_text|>' llm_load_print_meta: LF token = 128 'Ä' llm_load_print_meta: EOT token = 128009 '<|eot_id|>' 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 = 0.30 MiB llm_load_tensors: offloading 25 repeating layers to GPU llm_load_tensors: offloaded 25/33 layers to GPU llm_load_tensors: CPU buffer size = 30633.02 MiB llm_load_tensors: CUDA0 buffer size = 20800.78 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 = 500000.0 llama_new_context_with_model: freq_scale = 1 llama_kv_cache_init: CUDA_Host KV buffer size = 14.00 MiB llama_kv_cache_init: CUDA0 KV buffer size = 50.00 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 = 0.49 MiB llama_new_context_with_model: CUDA0 compute buffer size = 2262.50 MiB llama_new_context_with_model: CUDA_Host compute buffer size = 9.01 MiB llama_new_context_with_model: graph nodes = 1030 llama_new_context_with_model: graph splits = 81 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 68.266 ms compute_imatrix: computing over 189 chunks with batch_size 512 compute_imatrix: 0.99 seconds per pass - ETA 3.12 minutes [1]13.5054,[2]9.7458,[3]7.9960,[4]11.1449,[5]11.0617,[6]8.4871,[7]9.1466,[8]10.2816,[9]10.8623, save_imatrix: stored collected data after 10 chunks in Llama-3-Instruct-8B-SimPO-IMat-GGUF/imatrix.dat [10]10.8479,[11]11.9778,[12]11.7752,[13]12.7516,[14]13.8989,[15]14.6603,[16]15.7142,[17]17.3020,[18]17.6246,[19]16.2228, save_imatrix: stored collected data after 20 chunks in Llama-3-Instruct-8B-SimPO-IMat-GGUF/imatrix.dat [20]15.8390,[21]15.1440,[22]14.2869,[23]13.9109,[24]13.3731,[25]13.8954,[26]14.2643,[27]14.6002,[28]14.5233,[29]13.6040, save_imatrix: stored collected data after 30 chunks in Llama-3-Instruct-8B-SimPO-IMat-GGUF/imatrix.dat [30]12.9851,[31]12.7048,[32]12.7961,[33]12.8746,[34]13.0446,[35]13.6416,[36]13.9640,[37]14.6468,[38]14.8164,[39]15.3916, save_imatrix: stored collected data after 40 chunks in Llama-3-Instruct-8B-SimPO-IMat-GGUF/imatrix.dat [40]15.8105,[41]16.4036,[42]16.5336,[43]16.8982,[44]16.7492,[45]16.6613,[46]16.8116,[47]17.0547,[48]17.0528,[49]17.3634, save_imatrix: stored collected data after 50 chunks in Llama-3-Instruct-8B-SimPO-IMat-GGUF/imatrix.dat [50]17.6674,[51]17.8103,[52]17.7707,[53]17.8689,[54]17.8847,[55]17.7751,[56]17.6550,[57]17.6308,[58]17.7272,[59]17.9819, save_imatrix: stored collected data after 60 chunks in Llama-3-Instruct-8B-SimPO-IMat-GGUF/imatrix.dat [60]18.1656,[61]17.9078,[62]17.6227,[63]17.6967,[64]17.6428,[65]17.4757,[66]17.5360,[67]17.2861,[68]17.2680,[69]17.2024, save_imatrix: stored collected data after 70 chunks in Llama-3-Instruct-8B-SimPO-IMat-GGUF/imatrix.dat [70]17.3062,[71]17.5023,[72]17.5536,[73]17.5133,[74]17.5611,[75]17.2673,[76]17.1471,[77]16.9608,[78]16.9346,[79]16.8790, save_imatrix: stored collected data after 80 chunks in Llama-3-Instruct-8B-SimPO-IMat-GGUF/imatrix.dat [80]16.7729,[81]16.5190,[82]16.2986,[83]16.0855,[84]16.1458,[85]16.2517,[86]16.3207,[87]16.1732,[88]16.1114,[89]16.1689, save_imatrix: stored collected data after 90 chunks in Llama-3-Instruct-8B-SimPO-IMat-GGUF/imatrix.dat [90]16.2805,[91]16.2844,[92]16.2380,[93]16.3351,[94]16.3662,[95]16.2612,[96]16.3178,[97]16.3095,[98]16.2858,[99]16.2821, save_imatrix: stored collected data after 100 chunks in Llama-3-Instruct-8B-SimPO-IMat-GGUF/imatrix.dat [100]16.2498,[101]16.2234,[102]16.1577,[103]16.3242,[104]16.3959,[105]16.3712,[106]16.4666,[107]16.5587,[108]16.3406,[109]16.3208, save_imatrix: stored collected data after 110 chunks in Llama-3-Instruct-8B-SimPO-IMat-GGUF/imatrix.dat [110]16.2181,[111]16.1190,[112]16.0451,[113]15.8999,[114]15.7520,[115]15.5792,[116]15.4479,[117]15.3227,[118]15.2060,[119]15.2707, save_imatrix: stored collected data after 120 chunks in Llama-3-Instruct-8B-SimPO-IMat-GGUF/imatrix.dat [120]15.2302,[121]15.2141,[122]15.2753,[123]15.2914,[124]15.3559,[125]15.4386,[126]15.5440,[127]15.6451,[128]15.7717,[129]15.8865, save_imatrix: stored collected data after 130 chunks in Llama-3-Instruct-8B-SimPO-IMat-GGUF/imatrix.dat [130]15.7735,[131]15.7584,[132]15.7380,[133]15.7639,[134]15.6763,[135]15.6282,[136]15.6568,[137]15.6235,[138]15.5943,[139]15.6168, save_imatrix: stored collected data after 140 chunks in Llama-3-Instruct-8B-SimPO-IMat-GGUF/imatrix.dat [140]15.5926,[141]15.5378,[142]15.5153,[143]15.3972,[144]15.4808,[145]15.6099,[146]15.6820,[147]15.7546,[148]15.8333,[149]15.8874, save_imatrix: stored collected data after 150 chunks in Llama-3-Instruct-8B-SimPO-IMat-GGUF/imatrix.dat [150]15.8522,[151]15.8344,[152]15.8832,[153]16.0355,[154]16.1506,[155]16.0708,[156]16.0766,[157]16.1740,[158]16.2816,[159]16.4911, save_imatrix: stored collected data after 160 chunks in Llama-3-Instruct-8B-SimPO-IMat-GGUF/imatrix.dat [160]16.6219,[161]16.6934,[162]16.7543,[163]16.7444,[164]16.7619,[165]16.8787,[166]16.9058,[167]16.9057,[168]16.9796,[169]17.0488, save_imatrix: stored collected data after 170 chunks in Llama-3-Instruct-8B-SimPO-IMat-GGUF/imatrix.dat [170]17.0933,[171]17.1326,[172]17.1533,[173]17.1676,[174]17.1942,[175]17.1399,[176]17.1316,[177]17.1371,[178]17.1259,[179]17.0772, save_imatrix: stored collected data after 180 chunks in Llama-3-Instruct-8B-SimPO-IMat-GGUF/imatrix.dat [180]17.0066,[181]17.0970,[182]17.0098,[183]17.0803,[184]17.0968,[185]17.1787,[186]17.2275,[187]17.2125,[188]17.0941,[189]16.9891, save_imatrix: stored collected data after 189 chunks in Llama-3-Instruct-8B-SimPO-IMat-GGUF/imatrix.dat llama_print_timings: load time = 3041.13 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 = 169090.20 ms / 96768 tokens ( 1.75 ms per token, 572.29 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 = 172426.50 ms / 96769 tokens Final estimate: PPL = 16.9891 +/- 0.29030