main: build = 3006 (eaf6e031) main: built with cc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 for x86_64-linux-gnu main: seed = 1716798931 llama_model_loader: loaded meta data with 23 key-value pairs and 291 tensors from dolphin-2.9.1-llama-3-8b-IMat-GGUF/dolphin-2.9.1-llama-3-8b.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 = dolphin-2.9.1-llama-3-8b 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 = 1 llama_model_loader: - kv 11: llama.vocab_size u32 = 128258 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,128258] = ["!", "\"", "#", "$", "%", "&", "'", ... llama_model_loader: - kv 16: tokenizer.ggml.token_type arr[i32,128258] = [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 = 128256 llama_model_loader: - kv 20: tokenizer.ggml.padding_token_id u32 = 128001 llama_model_loader: - kv 21: tokenizer.chat_template str = {% if not add_generation_prompt is de... llama_model_loader: - kv 22: general.quantization_version u32 = 2 llama_model_loader: - type f32: 65 tensors llama_model_loader: - type f16: 226 tensors llm_load_vocab: special tokens definition check successful ( 258/128258 ). 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 = 128258 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 = F16 llm_load_print_meta: model params = 8.03 B llm_load_print_meta: model size = 14.96 GiB (16.00 BPW) llm_load_print_meta: general.name = dolphin-2.9.1-llama-3-8b llm_load_print_meta: BOS token = 128000 '<|begin_of_text|>' llm_load_print_meta: EOS token = 128256 '<|im_end|>' llm_load_print_meta: PAD token = 128001 '<|end_of_text|>' llm_load_print_meta: LF token = 128 'Ä' llm_load_print_meta: EOT token = 128256 '<|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 = 0.30 MiB llm_load_tensors: offloading 32 repeating layers to GPU llm_load_tensors: offloading non-repeating layers to GPU llm_load_tensors: offloaded 33/33 layers to GPU llm_load_tensors: CPU buffer size = 1002.02 MiB llm_load_tensors: CUDA0 buffer size = 14315.03 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: CUDA0 KV buffer size = 64.00 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 = 258.50 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 = 2 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 70.912 ms compute_imatrix: computing over 189 chunks with batch_size 512 compute_imatrix: 0.52 seconds per pass - ETA 1.63 minutes [1]9.1784,[2]7.1409,[3]6.1731,[4]7.8394,[5]8.1180,[6]6.6396,[7]7.2136,[8]8.2216,[9]8.3952, save_imatrix: stored collected data after 10 chunks in dolphin-2.9.1-llama-3-8b-IMat-GGUF/imatrix.dat [10]8.3728,[11]8.9763,[12]8.5482,[13]9.2517,[14]9.8158,[15]10.1556,[16]10.8281,[17]11.6711,[18]11.8423,[19]11.1913, save_imatrix: stored collected data after 20 chunks in dolphin-2.9.1-llama-3-8b-IMat-GGUF/imatrix.dat [20]11.0897,[21]10.5621,[22]9.7523,[23]9.4173,[24]9.1698,[25]9.5668,[26]9.8318,[27]10.0436,[28]10.1748,[29]9.7068, save_imatrix: stored collected data after 30 chunks in dolphin-2.9.1-llama-3-8b-IMat-GGUF/imatrix.dat [30]9.3447,[31]9.1863,[32]9.1298,[33]9.0470,[34]9.0370,[35]9.1829,[36]9.3919,[37]9.6569,[38]9.7511,[39]10.0391, save_imatrix: stored collected data after 40 chunks in dolphin-2.9.1-llama-3-8b-IMat-GGUF/imatrix.dat [40]10.3080,[41]10.6332,[42]10.8357,[43]11.1169,[44]11.0848,[45]11.0982,[46]11.2562,[47]11.4305,[48]11.4772,[49]11.5835, save_imatrix: stored collected data after 50 chunks in dolphin-2.9.1-llama-3-8b-IMat-GGUF/imatrix.dat [50]11.6733,[51]11.7400,[52]11.6936,[53]11.7463,[54]11.7177,[55]11.6926,[56]11.6243,[57]11.6318,[58]11.7222,[59]11.9595, save_imatrix: stored collected data after 60 chunks in dolphin-2.9.1-llama-3-8b-IMat-GGUF/imatrix.dat [60]12.0720,[61]11.9408,[62]11.8685,[63]11.7829,[64]11.7672,[65]11.6349,[66]11.5824,[67]11.4307,[68]11.4207,[69]11.3455, save_imatrix: stored collected data after 70 chunks in dolphin-2.9.1-llama-3-8b-IMat-GGUF/imatrix.dat [70]11.4202,[71]11.5000,[72]11.5165,[73]11.4903,[74]11.5322,[75]11.4093,[76]11.3549,[77]11.2360,[78]11.2357,[79]11.1787, save_imatrix: stored collected data after 80 chunks in dolphin-2.9.1-llama-3-8b-IMat-GGUF/imatrix.dat [80]11.1356,[81]11.1279,[82]11.0307,[83]10.9191,[84]10.9476,[85]11.0194,[86]11.0145,[87]10.9111,[88]10.9025,[89]10.9078, save_imatrix: stored collected data after 90 chunks in dolphin-2.9.1-llama-3-8b-IMat-GGUF/imatrix.dat [90]10.9639,[91]10.9442,[92]10.9349,[93]10.9655,[94]11.0145,[95]10.9527,[96]10.9654,[97]10.9747,[98]10.9568,[99]10.9686, save_imatrix: stored collected data after 100 chunks in dolphin-2.9.1-llama-3-8b-IMat-GGUF/imatrix.dat [100]10.9664,[101]10.9350,[102]10.9368,[103]10.9807,[104]11.0048,[105]10.9988,[106]11.0412,[107]11.0640,[108]10.9587,[109]10.9693, save_imatrix: stored collected data after 110 chunks in dolphin-2.9.1-llama-3-8b-IMat-GGUF/imatrix.dat [110]10.9356,[111]10.8382,[112]10.7911,[113]10.7187,[114]10.6528,[115]10.5696,[116]10.5105,[117]10.4490,[118]10.3922,[119]10.4675, save_imatrix: stored collected data after 120 chunks in dolphin-2.9.1-llama-3-8b-IMat-GGUF/imatrix.dat [120]10.4796,[121]10.5119,[122]10.5922,[123]10.6346,[124]10.7205,[125]10.8233,[126]10.9137,[127]11.0256,[128]11.1374,[129]11.2712, save_imatrix: stored collected data after 130 chunks in dolphin-2.9.1-llama-3-8b-IMat-GGUF/imatrix.dat [130]11.2120,[131]11.2382,[132]11.2576,[133]11.2884,[134]11.2568,[135]11.2486,[136]11.3074,[137]11.3059,[138]11.3240,[139]11.3672, save_imatrix: stored collected data after 140 chunks in dolphin-2.9.1-llama-3-8b-IMat-GGUF/imatrix.dat [140]11.3863,[141]11.3803,[142]11.4039,[143]11.3482,[144]11.3781,[145]11.4219,[146]11.4362,[147]11.4453,[148]11.4672,[149]11.4926, save_imatrix: stored collected data after 150 chunks in dolphin-2.9.1-llama-3-8b-IMat-GGUF/imatrix.dat [150]11.4594,[151]11.4308,[152]11.4498,[153]11.4699,[154]11.5420,[155]11.4944,[156]11.4899,[157]11.5502,[158]11.6104,[159]11.7480, save_imatrix: stored collected data after 160 chunks in dolphin-2.9.1-llama-3-8b-IMat-GGUF/imatrix.dat [160]11.8423,[161]11.8709,[162]11.8969,[163]11.9097,[164]11.9378,[165]11.9857,[166]11.9943,[167]11.9998,[168]12.0112,[169]12.0437, save_imatrix: stored collected data after 170 chunks in dolphin-2.9.1-llama-3-8b-IMat-GGUF/imatrix.dat [170]12.0301,[171]12.0139,[172]12.0363,[173]12.0320,[174]12.0483,[175]12.0209,[176]12.0146,[177]12.0165,[178]12.0192,[179]12.0088, save_imatrix: stored collected data after 180 chunks in dolphin-2.9.1-llama-3-8b-IMat-GGUF/imatrix.dat [180]11.9846,[181]12.0050,[182]11.9704,[183]12.0021,[184]12.0202,[185]12.0840,[186]12.1126,[187]12.1348,[188]12.0675,[189]12.0083, save_imatrix: stored collected data after 189 chunks in dolphin-2.9.1-llama-3-8b-IMat-GGUF/imatrix.dat llama_print_timings: load time = 2201.22 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 = 78396.70 ms / 96768 tokens ( 0.81 ms per token, 1234.34 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 = 81299.68 ms / 96769 tokens Final estimate: PPL = 12.0083 +/- 0.17709