main: build = 3003 (d298382a) main: built with cc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 for x86_64-linux-gnu main: seed = 1716764546 llama_model_loader: loaded meta data with 23 key-value pairs and 291 tensors from Hermes-2-Theta-Llama-3-8B-IMat-GGUF/Hermes-2-Theta-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 = Hermes-2-Theta-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 = 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 = 128003 llama_model_loader: - kv 20: tokenizer.ggml.padding_token_id u32 = 128001 llama_model_loader: - kv 21: tokenizer.chat_template str = {{bos_token}}{% for message in messag... llama_model_loader: - kv 22: 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 = Hermes-2-Theta-Llama-3-8B llm_load_print_meta: BOS token = 128000 '<|begin_of_text|>' llm_load_print_meta: EOS token = 128003 '<|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 = 128003 '<|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 23 repeating layers to GPU llm_load_tensors: offloaded 23/33 layers to GPU llm_load_tensors: CPU buffer size = 30633.02 MiB llm_load_tensors: CUDA0 buffer size = 19136.72 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 = 18.00 MiB llama_kv_cache_init: CUDA0 KV buffer size = 46.00 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 = 103 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 71.123 ms compute_imatrix: computing over 189 chunks with batch_size 512 compute_imatrix: 1.10 seconds per pass - ETA 3.47 minutes [1]6.7093,[2]5.2192,[3]4.6281,[4]5.8521,[5]5.8939,[6]4.9505,[7]5.2875,[8]5.8677,[9]6.0817, save_imatrix: stored collected data after 10 chunks in Hermes-2-Theta-Llama-3-8B-IMat-GGUF/imatrix.dat [10]6.0511,[11]6.5621,[12]6.3089,[13]6.7770,[14]7.2232,[15]7.4788,[16]7.9077,[17]8.3941,[18]8.5694,[19]8.1400, save_imatrix: stored collected data after 20 chunks in Hermes-2-Theta-Llama-3-8B-IMat-GGUF/imatrix.dat [20]8.0283,[21]7.7469,[22]7.2594,[23]6.9051,[24]6.6862,[25]6.9238,[26]7.0536,[27]7.2167,[28]7.2089,[29]6.8810, save_imatrix: stored collected data after 30 chunks in Hermes-2-Theta-Llama-3-8B-IMat-GGUF/imatrix.dat [30]6.6645,[31]6.5576,[32]6.5316,[33]6.5106,[34]6.5147,[35]6.6542,[36]6.7603,[37]6.9338,[38]7.0092,[39]7.1740, save_imatrix: stored collected data after 40 chunks in Hermes-2-Theta-Llama-3-8B-IMat-GGUF/imatrix.dat [40]7.3507,[41]7.5822,[42]7.7183,[43]7.8980,[44]7.8788,[45]7.8878,[46]7.9950,[47]8.1496,[48]8.1852,[49]8.2834, save_imatrix: stored collected data after 50 chunks in Hermes-2-Theta-Llama-3-8B-IMat-GGUF/imatrix.dat [50]8.3087,[51]8.3517,[52]8.2651,[53]8.2901,[54]8.2892,[55]8.1896,[56]8.0818,[57]8.0914,[58]8.1647,[59]8.2764, save_imatrix: stored collected data after 60 chunks in Hermes-2-Theta-Llama-3-8B-IMat-GGUF/imatrix.dat [60]8.3593,[61]8.2742,[62]8.1604,[63]8.0614,[64]8.0129,[65]7.9229,[66]7.8169,[67]7.6771,[68]7.6417,[69]7.5732, save_imatrix: stored collected data after 70 chunks in Hermes-2-Theta-Llama-3-8B-IMat-GGUF/imatrix.dat [70]7.5909,[71]7.6555,[72]7.6823,[73]7.6835,[74]7.7239,[75]7.6178,[76]7.4781,[77]7.3407,[78]7.2739,[79]7.2243, save_imatrix: stored collected data after 80 chunks in Hermes-2-Theta-Llama-3-8B-IMat-GGUF/imatrix.dat [80]7.1912,[81]7.0820,[82]7.0080,[83]6.9500,[84]6.9788,[85]7.0240,[86]7.0268,[87]6.9794,[88]6.9724,[89]6.9929, save_imatrix: stored collected data after 90 chunks in Hermes-2-Theta-Llama-3-8B-IMat-GGUF/imatrix.dat [90]7.0337,[91]7.0293,[92]7.0346,[93]7.0664,[94]7.1015,[95]7.0799,[96]7.1100,[97]7.1220,[98]7.1252,[99]7.1422, save_imatrix: stored collected data after 100 chunks in Hermes-2-Theta-Llama-3-8B-IMat-GGUF/imatrix.dat [100]7.1407,[101]7.1314,[102]7.1332,[103]7.1716,[104]7.2006,[105]7.2004,[106]7.2386,[107]7.2735,[108]7.2128,[109]7.2214, save_imatrix: stored collected data after 110 chunks in Hermes-2-Theta-Llama-3-8B-IMat-GGUF/imatrix.dat [110]7.2009,[111]7.1498,[112]7.1247,[113]7.0841,[114]7.0388,[115]6.9946,[116]6.9513,[117]6.9099,[118]6.8721,[119]6.9224, save_imatrix: stored collected data after 120 chunks in Hermes-2-Theta-Llama-3-8B-IMat-GGUF/imatrix.dat [120]6.9365,[121]6.9595,[122]7.0130,[123]7.0512,[124]7.1134,[125]7.1760,[126]7.2420,[127]7.3016,[128]7.3830,[129]7.4697, save_imatrix: stored collected data after 130 chunks in Hermes-2-Theta-Llama-3-8B-IMat-GGUF/imatrix.dat [130]7.4394,[131]7.4658,[132]7.4830,[133]7.5096,[134]7.4936,[135]7.4930,[136]7.5370,[137]7.5490,[138]7.5630,[139]7.5897, save_imatrix: stored collected data after 140 chunks in Hermes-2-Theta-Llama-3-8B-IMat-GGUF/imatrix.dat [140]7.6117,[141]7.6153,[142]7.6337,[143]7.5992,[144]7.6183,[145]7.6505,[146]7.6661,[147]7.6744,[148]7.6892,[149]7.7072, save_imatrix: stored collected data after 150 chunks in Hermes-2-Theta-Llama-3-8B-IMat-GGUF/imatrix.dat [150]7.6860,[151]7.6757,[152]7.6887,[153]7.7072,[154]7.7552,[155]7.7220,[156]7.7235,[157]7.7649,[158]7.8167,[159]7.9031, save_imatrix: stored collected data after 160 chunks in Hermes-2-Theta-Llama-3-8B-IMat-GGUF/imatrix.dat [160]7.9716,[161]7.9894,[162]8.0062,[163]8.0187,[164]8.0259,[165]8.0608,[166]8.0651,[167]8.0671,[168]8.0833,[169]8.1123, save_imatrix: stored collected data after 170 chunks in Hermes-2-Theta-Llama-3-8B-IMat-GGUF/imatrix.dat [170]8.1154,[171]8.1165,[172]8.1331,[173]8.1187,[174]8.1275,[175]8.1136,[176]8.1099,[177]8.1108,[178]8.1102,[179]8.1024, save_imatrix: stored collected data after 180 chunks in Hermes-2-Theta-Llama-3-8B-IMat-GGUF/imatrix.dat [180]8.0850,[181]8.1010,[182]8.0732,[183]8.0683,[184]8.0413,[185]8.0746,[186]8.0759,[187]8.0723,[188]8.0278,[189]7.9888, save_imatrix: stored collected data after 189 chunks in Hermes-2-Theta-Llama-3-8B-IMat-GGUF/imatrix.dat llama_print_timings: load time = 3066.66 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 = 184880.42 ms / 96768 tokens ( 1.91 ms per token, 523.41 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 = 188125.37 ms / 96769 tokens Final estimate: PPL = 7.9888 +/- 0.10218