llama_model_loader: loaded meta data with 30 key-value pairs and 291 tensors from internlm2_5-7b-chat-IMat-GGUF/internlm2_5-7b-chat.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 = internlm2 llama_model_loader: - kv 1: general.type str = model llama_model_loader: - kv 2: general.name str = Internlm2_5 7b Chat llama_model_loader: - kv 3: general.finetune str = chat llama_model_loader: - kv 4: general.basename str = internlm2_5 llama_model_loader: - kv 5: general.size_label str = 7B llama_model_loader: - kv 6: general.license str = other llama_model_loader: - kv 7: general.tags arr[str,1] = ["text-generation"] llama_model_loader: - kv 8: internlm2.context_length u32 = 32768 llama_model_loader: - kv 9: internlm2.block_count u32 = 32 llama_model_loader: - kv 10: internlm2.embedding_length u32 = 4096 llama_model_loader: - kv 11: internlm2.feed_forward_length u32 = 14336 llama_model_loader: - kv 12: internlm2.rope.freq_base f32 = 1000000.000000 llama_model_loader: - kv 13: internlm2.attention.head_count u32 = 32 llama_model_loader: - kv 14: internlm2.attention.layer_norm_rms_epsilon f32 = 0.000010 llama_model_loader: - kv 15: internlm2.attention.head_count_kv u32 = 8 llama_model_loader: - kv 16: general.file_type u32 = 7 llama_model_loader: - kv 17: tokenizer.ggml.model str = llama llama_model_loader: - kv 18: tokenizer.ggml.pre str = default llama_model_loader: - kv 19: tokenizer.ggml.tokens arr[str,92544] = ["", "~~", "~~", "<0x00>", "<... llama_model_loader: - kv 20: tokenizer.ggml.scores arr[f32,92544] = [-1000.000000, -1000.000000, -1000.00... llama_model_loader: - kv 21: tokenizer.ggml.token_type arr[i32,92544] = [3, 3, 3, 6, 6, 6, 6, 6, 6, 6, 6, 6, ... llama_model_loader: - kv 22: tokenizer.ggml.add_space_prefix bool = false llama_model_loader: - kv 23: tokenizer.ggml.bos_token_id u32 = 1 llama_model_loader: - kv 24: tokenizer.ggml.eos_token_id u32 = 92542 llama_model_loader: - kv 25: tokenizer.ggml.padding_token_id u32 = 2 llama_model_loader: - kv 26: tokenizer.ggml.add_bos_token bool = true llama_model_loader: - kv 27: tokenizer.ggml.add_eos_token bool = false llama_model_loader: - kv 28: tokenizer.chat_template str = {{ bos_token }}{% for message in mess... llama_model_loader: - kv 29: general.quantization_version u32 = 2 llama_model_loader: - type f32: 65 tensors llama_model_loader: - type q8_0: 226 tensors llm_load_vocab: special tokens cache size = 9 llm_load_vocab: token to piece cache size = 0.5508 MB llm_load_print_meta: format = GGUF V3 (latest) llm_load_print_meta: arch = internlm2 llm_load_print_meta: vocab type = SPM llm_load_print_meta: n_vocab = 92544 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 = 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 = 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: model type = 7B llm_load_print_meta: model ftype = Q8_0 llm_load_print_meta: model params = 7.74 B llm_load_print_meta: model size = 7.66 GiB (8.50 BPW) llm_load_print_meta: general.name = Internlm2_5 7b Chat llm_load_print_meta: BOS token = 1 '~~' llm_load_print_meta: EOS token = 92542 '<|im_end|>' llm_load_print_meta: UNK token = 0 '' llm_load_print_meta: PAD token = 2 '~~' llm_load_print_meta: LF token = 13 '<0x0A>' llm_load_print_meta: EOT token = 92542 '<|im_end|>' llm_load_print_meta: max token length = 384 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 = 0.27 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 = 384.09 MiB llm_load_tensors: CUDA0 buffer size = 7457.11 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: 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 = 0.35 MiB llama_new_context_with_model: CUDA0 compute buffer size = 188.75 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 101.295 ms compute_imatrix: computing over 136 chunks with batch_size 512 compute_imatrix: 0.59 seconds per pass - ETA 1.32 minutes [1]5.3769,[2]3.9691,[3]3.7820,[4]4.3833,[5]4.2913,[6]3.9200,[7]4.5687,[8]4.6228,[9]5.0904, save_imatrix: stored collected data after 10 chunks in internlm2_5-7b-chat-IMat-GGUF/imatrix.dat [10]5.2358,[11]4.8009,[12]5.1499,[13]5.7993,[14]6.1283,[15]6.6242,[16]6.8570,[17]6.5016,[18]6.7125,[19]7.0479, save_imatrix: stored collected data after 20 chunks in internlm2_5-7b-chat-IMat-GGUF/imatrix.dat [20]6.8082,[21]6.8744,[22]7.0183,[23]7.0556,[24]7.0649,[25]7.2574,[26]7.4557,[27]7.6284,[28]7.6740,[29]7.8141, save_imatrix: stored collected data after 30 chunks in internlm2_5-7b-chat-IMat-GGUF/imatrix.dat [30]8.0461,[31]8.1025,[32]7.8055,[33]7.4904,[34]7.1984,[35]6.9328,[36]6.7832,[37]6.6670,[38]6.5876,[39]6.5080, save_imatrix: stored collected data after 40 chunks in internlm2_5-7b-chat-IMat-GGUF/imatrix.dat [40]6.4166,[41]6.3603,[42]6.2501,[43]6.2315,[44]6.2919,[45]6.3633,[46]6.4997,[47]6.4768,[48]6.7117,[49]6.8956, save_imatrix: stored collected data after 50 chunks in internlm2_5-7b-chat-IMat-GGUF/imatrix.dat [50]7.0720,[51]7.2003,[52]7.3871,[53]7.2634,[54]7.3487,[55]7.4317,[56]7.5508,[57]7.4196,[58]7.4276,[59]7.4583, save_imatrix: stored collected data after 60 chunks in internlm2_5-7b-chat-IMat-GGUF/imatrix.dat [60]7.5612,[61]7.6982,[62]7.8157,[63]7.8791,[64]7.8829,[65]7.8916,[66]7.8631,[67]7.8273,[68]7.7453,[69]7.7176, save_imatrix: stored collected data after 70 chunks in internlm2_5-7b-chat-IMat-GGUF/imatrix.dat [70]7.7884,[71]7.8087,[72]7.7333,[73]7.6979,[74]7.7003,[75]7.6571,[76]7.6477,[77]7.6271,[78]7.6449,[79]7.5887, save_imatrix: stored collected data after 80 chunks in internlm2_5-7b-chat-IMat-GGUF/imatrix.dat [80]7.5901,[81]7.5365,[82]7.5011,[83]7.4550,[84]7.4343,[85]7.3733,[86]7.3403,[87]7.3036,[88]7.3420,[89]7.3542, save_imatrix: stored collected data after 90 chunks in internlm2_5-7b-chat-IMat-GGUF/imatrix.dat [90]7.3164,[91]7.3376,[92]7.3530,[93]7.3026,[94]7.2918,[95]7.2765,[96]7.3050,[97]7.3023,[98]7.3053,[99]7.2564, save_imatrix: stored collected data after 100 chunks in internlm2_5-7b-chat-IMat-GGUF/imatrix.dat [100]7.2231,[101]7.1735,[102]7.1225,[103]7.0841,[104]7.0417,[105]7.0015,[106]6.9774,[107]6.9807,[108]7.0191,[109]7.0920, save_imatrix: stored collected data after 110 chunks in internlm2_5-7b-chat-IMat-GGUF/imatrix.dat [110]7.1641,[111]7.2266,[112]7.3301,[113]7.3983,[114]7.4227,[115]7.4135,[116]7.4296,[117]7.4244,[118]7.4193,[119]7.3785, save_imatrix: stored collected data after 120 chunks in internlm2_5-7b-chat-IMat-GGUF/imatrix.dat [120]7.3659,[121]7.3952,[122]7.3990,[123]7.4023,[124]7.4202,[125]7.4587,[126]7.4862,[127]7.4988,[128]7.5187,[129]7.5457, save_imatrix: stored collected data after 130 chunks in internlm2_5-7b-chat-IMat-GGUF/imatrix.dat [130]7.5014,[131]7.5428,[132]7.6069,[133]7.6466,[134]7.7078,[135]7.7539,[136]7.8028, save_imatrix: stored collected data after 136 chunks in internlm2_5-7b-chat-IMat-GGUF/imatrix.dat llama_print_timings: load time = 2096.79 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 = 70631.81 ms / 69632 tokens ( 1.01 ms per token, 985.84 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 = 72872.89 ms / 69633 tokens Final estimate: PPL = 7.8028 +/- 0.10727