File size: 12,247 Bytes
f2b6c73 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 |
llama_model_loader: loaded meta data with 32 key-value pairs and 628 tensors from granite-20b-code-instruct-IMat-GGUF/granite-20b-code-instruct.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 = starcoder
llama_model_loader: - kv 1: general.type str = model
llama_model_loader: - kv 2: general.name str = Granite 20b Code Instruct
llama_model_loader: - kv 3: general.finetune str = code-instruct
llama_model_loader: - kv 4: general.basename str = granite
llama_model_loader: - kv 5: general.size_label str = 20B
llama_model_loader: - kv 6: general.license str = apache-2.0
llama_model_loader: - kv 7: general.base_model.count u32 = 1
llama_model_loader: - kv 8: general.base_model.0.name str = Granite 20b Code Base
llama_model_loader: - kv 9: general.base_model.0.organization str = Ibm Granite
llama_model_loader: - kv 10: general.base_model.0.repo_url str = https://huggingface.co/ibm-granite/gr...
llama_model_loader: - kv 11: general.tags arr[str,3] = ["code", "granite", "text-generation"]
llama_model_loader: - kv 12: general.datasets arr[str,8] = ["bigcode/commitpackft", "TIGER-Lab/M...
llama_model_loader: - kv 13: starcoder.context_length u32 = 8192
llama_model_loader: - kv 14: starcoder.embedding_length u32 = 6144
llama_model_loader: - kv 15: starcoder.feed_forward_length u32 = 24576
llama_model_loader: - kv 16: starcoder.block_count u32 = 52
llama_model_loader: - kv 17: starcoder.attention.head_count u32 = 48
llama_model_loader: - kv 18: starcoder.attention.head_count_kv u32 = 1
llama_model_loader: - kv 19: starcoder.attention.layer_norm_epsilon f32 = 0.000010
llama_model_loader: - kv 20: general.file_type u32 = 7
llama_model_loader: - kv 21: tokenizer.ggml.model str = gpt2
llama_model_loader: - kv 22: tokenizer.ggml.pre str = refact
llama_model_loader: - kv 23: tokenizer.ggml.tokens arr[str,49152] = ["<|endoftext|>", "<fim_prefix>", "<f...
llama_model_loader: - kv 24: tokenizer.ggml.token_type arr[i32,49152] = [3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, ...
llama_model_loader: - kv 25: tokenizer.ggml.merges arr[str,48891] = ["Ġ Ġ", "ĠĠ ĠĠ", "ĠĠĠĠ ĠĠ...
llama_model_loader: - kv 26: tokenizer.ggml.bos_token_id u32 = 0
llama_model_loader: - kv 27: tokenizer.ggml.eos_token_id u32 = 0
llama_model_loader: - kv 28: tokenizer.ggml.unknown_token_id u32 = 0
llama_model_loader: - kv 29: tokenizer.ggml.padding_token_id u32 = 0
llama_model_loader: - kv 30: tokenizer.chat_template str = {% for message in messages %}\n{% if m...
llama_model_loader: - kv 31: general.quantization_version u32 = 2
llama_model_loader: - type f32: 419 tensors
llama_model_loader: - type q8_0: 209 tensors
llm_load_vocab: special tokens cache size = 19
llm_load_vocab: token to piece cache size = 0.2826 MB
llm_load_print_meta: format = GGUF V3 (latest)
llm_load_print_meta: arch = starcoder
llm_load_print_meta: vocab type = BPE
llm_load_print_meta: n_vocab = 49152
llm_load_print_meta: n_merges = 48891
llm_load_print_meta: vocab_only = 0
llm_load_print_meta: n_ctx_train = 8192
llm_load_print_meta: n_embd = 6144
llm_load_print_meta: n_layer = 52
llm_load_print_meta: n_head = 48
llm_load_print_meta: n_head_kv = 1
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 = 48
llm_load_print_meta: n_embd_k_gqa = 128
llm_load_print_meta: n_embd_v_gqa = 128
llm_load_print_meta: f_norm_eps = 1.0e-05
llm_load_print_meta: f_norm_rms_eps = 0.0e+00
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 = 24576
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 = 10000.0
llm_load_print_meta: freq_scale_train = 1
llm_load_print_meta: n_ctx_orig_yarn = 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: 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 = 20.07 B
llm_load_print_meta: model size = 20.00 GiB (8.56 BPW)
llm_load_print_meta: general.name = Granite 20b Code Instruct
llm_load_print_meta: BOS token = 0 '<|endoftext|>'
llm_load_print_meta: EOS token = 0 '<|endoftext|>'
llm_load_print_meta: UNK token = 0 '<|endoftext|>'
llm_load_print_meta: PAD token = 0 '<|endoftext|>'
llm_load_print_meta: LF token = 145 'Ä'
llm_load_print_meta: EOT token = 0 '<|endoftext|>'
llm_load_print_meta: max token length = 512
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 = 0.55 MiB
llm_load_tensors: offloading 52 repeating layers to GPU
llm_load_tensors: offloading non-repeating layers to GPU
llm_load_tensors: offloaded 53/53 layers to GPU
llm_load_tensors: CPU buffer size = 498.00 MiB
llm_load_tensors: CUDA0 buffer size = 20292.38 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 = 10000.0
llama_new_context_with_model: freq_scale = 1
llama_kv_cache_init: CUDA0 KV buffer size = 13.00 MiB
llama_new_context_with_model: KV self size = 13.00 MiB, K (f16): 6.50 MiB, V (f16): 6.50 MiB
llama_new_context_with_model: CUDA_Host output buffer size = 0.19 MiB
llama_new_context_with_model: CUDA0 compute buffer size = 120.00 MiB
llama_new_context_with_model: CUDA_Host compute buffer size = 25.01 MiB
llama_new_context_with_model: graph nodes = 1933
llama_new_context_with_model: graph splits = 2
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 | 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 113.128 ms
compute_imatrix: computing over 152 chunks with batch_size 512
compute_imatrix: 1.30 seconds per pass - ETA 3.28 minutes
[1]5.5316,[2]4.3038,[3]4.5649,[4]5.0288,[5]5.6039,[6]5.7152,[7]5.0016,[8]5.9597,[9]5.9637,
save_imatrix: stored collected data after 10 chunks in granite-20b-code-instruct-IMat-GGUF/imatrix.dat
[10]6.7819,[11]7.0343,[12]6.4546,[13]7.0672,[14]7.6034,[15]8.3912,[16]8.5493,[17]9.1657,[18]9.4702,[19]9.7523,
save_imatrix: stored collected data after 20 chunks in granite-20b-code-instruct-IMat-GGUF/imatrix.dat
[20]9.9486,[21]10.4114,[22]10.0740,[23]10.0213,[24]10.3530,[25]10.5995,[26]10.6331,[27]10.4010,[28]10.5028,[29]10.7926,
save_imatrix: stored collected data after 30 chunks in granite-20b-code-instruct-IMat-GGUF/imatrix.dat
[30]11.0959,[31]11.0268,[32]11.1389,[33]11.3398,[34]11.6580,[35]11.7255,[36]11.4972,[37]10.8076,[38]10.3152,[39]10.2085,
save_imatrix: stored collected data after 40 chunks in granite-20b-code-instruct-IMat-GGUF/imatrix.dat
[40]10.1415,[41]10.0520,[42]10.1053,[43]9.9861,[44]10.0245,[45]10.0257,[46]10.1471,[47]10.1358,[48]10.3092,[49]10.5322,
save_imatrix: stored collected data after 50 chunks in granite-20b-code-instruct-IMat-GGUF/imatrix.dat
[50]10.4987,[51]10.8835,[52]11.2237,[53]11.6686,[54]11.9040,[55]12.2365,[56]12.2093,[57]12.1756,[58]12.1845,[59]12.3144,
save_imatrix: stored collected data after 60 chunks in granite-20b-code-instruct-IMat-GGUF/imatrix.dat
[60]12.3865,[61]12.2209,[62]12.2149,[63]12.2819,[64]12.3703,[65]12.4174,[66]12.4922,[67]12.6439,[68]12.7639,[69]12.7716,
save_imatrix: stored collected data after 70 chunks in granite-20b-code-instruct-IMat-GGUF/imatrix.dat
[70]12.8490,[71]12.8158,[72]12.6971,[73]12.6002,[74]12.4168,[75]12.3231,[76]12.4371,[77]12.5235,[78]12.4813,[79]12.4734,
save_imatrix: stored collected data after 80 chunks in granite-20b-code-instruct-IMat-GGUF/imatrix.dat
[80]12.4798,[81]12.3699,[82]12.2741,[83]12.2127,[84]12.1877,[85]12.1598,[86]12.0773,[87]12.0180,[88]11.9234,[89]11.8501,
save_imatrix: stored collected data after 90 chunks in granite-20b-code-instruct-IMat-GGUF/imatrix.dat
[90]11.7759,[91]11.7479,[92]11.6583,[93]11.5726,[94]11.4962,[95]11.4289,[96]11.4031,[97]11.3468,[98]11.3321,[99]11.2733,
save_imatrix: stored collected data after 100 chunks in granite-20b-code-instruct-IMat-GGUF/imatrix.dat
[100]11.2175,[101]11.2073,[102]11.1154,[103]11.0428,[104]11.0297,[105]11.0836,[106]11.1179,[107]11.1815,[108]11.2633,[109]11.1341,
save_imatrix: stored collected data after 110 chunks in granite-20b-code-instruct-IMat-GGUF/imatrix.dat
[110]11.0276,[111]10.9195,[112]10.7971,[113]10.6767,[114]10.5720,[115]10.4712,[116]10.3728,[117]10.3827,[118]10.3708,[119]10.3970,
save_imatrix: stored collected data after 120 chunks in granite-20b-code-instruct-IMat-GGUF/imatrix.dat
[120]10.5022,[121]10.6177,[122]10.7352,[123]10.7968,[124]10.9187,[125]10.9587,[126]10.9877,[127]11.0106,[128]11.0613,[129]11.0879,
save_imatrix: stored collected data after 130 chunks in granite-20b-code-instruct-IMat-GGUF/imatrix.dat
[130]11.0612,[131]11.0532,[132]11.0083,[133]11.0206,[134]11.0336,[135]11.0436,[136]11.0618,[137]11.0628,[138]11.0782,[139]11.0803,
save_imatrix: stored collected data after 140 chunks in granite-20b-code-instruct-IMat-GGUF/imatrix.dat
[140]11.1238,[141]11.1233,[142]11.1360,[143]11.1245,[144]11.0680,[145]11.0480,[146]11.1301,[147]11.2386,[148]11.3118,[149]11.4027,
save_imatrix: stored collected data after 150 chunks in granite-20b-code-instruct-IMat-GGUF/imatrix.dat
[150]11.4866,[151]11.5899,[152]11.6361,
save_imatrix: stored collected data after 152 chunks in granite-20b-code-instruct-IMat-GGUF/imatrix.dat
llama_print_timings: load time = 11508.22 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 = 173482.28 ms / 77824 tokens ( 2.23 ms per token, 448.60 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 = 184251.37 ms / 77825 tokens
Final estimate: PPL = 11.6361 +/- 0.19854
|