-
Notifications
You must be signed in to change notification settings - Fork 0
/
main.log
102 lines (102 loc) · 9.47 KB
/
main.log
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
[1709429536] Log start
[1709429536] Cmd: /Users/crimsonknight/llama.cpp/main -m /Users/crimsonknight/.agentc/models/llama-2-13b-chat.Q8_0.gguf --grammar-file /Users/crimsonknight/.agentc/grammars/goal_planning.gbnf --n-predict 256 --in-suffix <Assistant: --threads 32 --ctx-size 2048 --temp 0.9 --top-k 64 --repeat-penalty 1.2 --prompt "You are a Senior Ruby on Rails developer. Write the first 20 commands you would enter into the CLI to create a new Rails app called 'Helpful Husband'. This app is a To Do List Management app that will help a husband and wife to manage a list of tasks and the completion status of the items"
[1709429536] main: build = 2261 (f1a98c52)
[1709429536] main: built with Apple clang version 15.0.0 (clang-1500.1.0.2.5) for arm64-apple-darwin23.2.0
[1709429536] main: seed = 1709429536
[1709429536] main: llama backend init
[1709429536] main: load the model and apply lora adapter, if any
[1709429536] llama_model_loader: loaded meta data with 19 key-value pairs and 363 tensors from /Users/crimsonknight/.agentc/models/llama-2-13b-chat.Q8_0.gguf (version GGUF V2)
[1709429536] llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
[1709429536] llama_model_loader: - kv 0: general.architecture str = llama
[1709429536] llama_model_loader: - kv 1: general.name str = LLaMA v2
[1709429536] llama_model_loader: - kv 2: llama.context_length u32 = 4096
[1709429536] llama_model_loader: - kv 3: llama.embedding_length u32 = 5120
[1709429536] llama_model_loader: - kv 4: llama.block_count u32 = 40
[1709429536] llama_model_loader: - kv 5: llama.feed_forward_length u32 = 13824
[1709429536] llama_model_loader: - kv 6: llama.rope.dimension_count u32 = 128
[1709429536] llama_model_loader: - kv 7: llama.attention.head_count u32 = 40
[1709429536] llama_model_loader: - kv 8: llama.attention.head_count_kv u32 = 40
[1709429536] llama_model_loader: - kv 9: llama.attention.layer_norm_rms_epsilon f32 = 0.000010
[1709429536] llama_model_loader: - kv 10: general.file_type u32 = 7
[1709429536] llama_model_loader: - kv 11: tokenizer.ggml.model str = llama
[1709429536] llama_model_loader: - kv 12: tokenizer.ggml.tokens arr[str,32000] = ["<unk>", "<s>", "</s>", "<0x00>", "<...
[1709429536] llama_model_loader: - kv 13: tokenizer.ggml.scores arr[f32,32000] = [0.000000, 0.000000, 0.000000, 0.0000...
[1709429536] llama_model_loader: - kv 14: tokenizer.ggml.token_type arr[i32,32000] = [2, 3, 3, 6, 6, 6, 6, 6, 6, 6, 6, 6, ...
[1709429536] llama_model_loader: - kv 15: tokenizer.ggml.bos_token_id u32 = 1
[1709429536] llama_model_loader: - kv 16: tokenizer.ggml.eos_token_id u32 = 2
[1709429536] llama_model_loader: - kv 17: tokenizer.ggml.unknown_token_id u32 = 0
[1709429536] llama_model_loader: - kv 18: general.quantization_version u32 = 2
[1709429536] llama_model_loader: - type f32: 81 tensors
[1709429536] llama_model_loader: - type q8_0: 282 tensors
[1709429536] llm_load_vocab: special tokens definition check successful ( 259/32000 ).
[1709429536] llm_load_print_meta: format = GGUF V2
[1709429536] llm_load_print_meta: arch = llama
[1709429536] llm_load_print_meta: vocab type = SPM
[1709429536] llm_load_print_meta: n_vocab = 32000
[1709429536] llm_load_print_meta: n_merges = 0
[1709429536] llm_load_print_meta: n_ctx_train = 4096
[1709429536] llm_load_print_meta: n_embd = 5120
[1709429536] llm_load_print_meta: n_head = 40
[1709429536] llm_load_print_meta: n_head_kv = 40
[1709429536] llm_load_print_meta: n_layer = 40
[1709429536] llm_load_print_meta: n_rot = 128
[1709429536] llm_load_print_meta: n_embd_head_k = 128
[1709429536] llm_load_print_meta: n_embd_head_v = 128
[1709429536] llm_load_print_meta: n_gqa = 1
[1709429536] llm_load_print_meta: n_embd_k_gqa = 5120
[1709429536] llm_load_print_meta: n_embd_v_gqa = 5120
[1709429536] llm_load_print_meta: f_norm_eps = 0.0e+00
[1709429536] llm_load_print_meta: f_norm_rms_eps = 1.0e-05
[1709429536] llm_load_print_meta: f_clamp_kqv = 0.0e+00
[1709429536] llm_load_print_meta: f_max_alibi_bias = 0.0e+00
[1709429536] llm_load_print_meta: n_ff = 13824
[1709429536] llm_load_print_meta: n_expert = 0
[1709429536] llm_load_print_meta: n_expert_used = 0
[1709429536] llm_load_print_meta: rope scaling = linear
[1709429536] llm_load_print_meta: freq_base_train = 10000.0
[1709429536] llm_load_print_meta: freq_scale_train = 1
[1709429536] llm_load_print_meta: n_yarn_orig_ctx = 4096
[1709429536] llm_load_print_meta: rope_finetuned = unknown
[1709429536] llm_load_print_meta: model type = 13B
[1709429536] llm_load_print_meta: model ftype = Q8_0
[1709429536] llm_load_print_meta: model params = 13.02 B
[1709429536] llm_load_print_meta: model size = 12.88 GiB (8.50 BPW)
[1709429536] llm_load_print_meta: general.name = LLaMA v2
[1709429536] llm_load_print_meta: BOS token = 1 '<s>'
[1709429536] llm_load_print_meta: EOS token = 2 '</s>'
[1709429536] llm_load_print_meta: UNK token = 0 '<unk>'
[1709429536] llm_load_print_meta: LF token = 13 '<0x0A>'
[1709429536] llm_load_tensors: ggml ctx size = 0.28 MiB
[1709429536] ggml_backend_metal_buffer_from_ptr: allocated buffer, size = 13023.86 MiB[1709429536] , (13023.92 / 21845.34)[1709429536]
[1709429536] llm_load_tensors: offloading 40 repeating layers to GPU
[1709429536] llm_load_tensors: offloading non-repeating layers to GPU
[1709429536] llm_load_tensors: offloaded 41/41 layers to GPU
[1709429536] llm_load_tensors: Metal buffer size = 13023.86 MiB
[1709429536] llm_load_tensors: CPU buffer size = 166.02 MiB
[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536] .[1709429536]
[1709429536] llama_new_context_with_model: n_ctx = 2048
[1709429536] llama_new_context_with_model: freq_base = 10000.0
[1709429536] llama_new_context_with_model: freq_scale = 1
[1709429536] ggml_metal_init: allocating
[1709429536] ggml_metal_init: found device: Apple M1 Max
[1709429536] ggml_metal_init: picking default device: Apple M1 Max
[1709429536] ggml_metal_init: default.metallib not found, loading from source
[1709429536] ggml_metal_init: GGML_METAL_PATH_RESOURCES = nil
[1709429536] ggml_metal_init: loading '/Users/crimsonknight/llama.cpp/ggml-metal.metal'
[1709429536] ggml_metal_init: GPU name: Apple M1 Max
[1709429536] ggml_metal_init: GPU family: MTLGPUFamilyApple7 (1007)
[1709429536] ggml_metal_init: GPU family: MTLGPUFamilyCommon3 (3003)
[1709429536] ggml_metal_init: GPU family: MTLGPUFamilyMetal3 (5001)
[1709429536] ggml_metal_init: simdgroup reduction support = true
[1709429536] ggml_metal_init: simdgroup matrix mul. support = true
[1709429536] ggml_metal_init: hasUnifiedMemory = true
[1709429536] ggml_metal_init: recommendedMaxWorkingSetSize = 22906.50 MB
[1709429536] ggml_backend_metal_buffer_type_alloc_buffer: allocated buffer, size = 1600.00 MiB[1709429536] , (14625.73 / 21845.34)[1709429536]
[1709429537] llama_kv_cache_init: Metal KV buffer size = 1600.00 MiB
[1709429537] llama_new_context_with_model: KV self size = 1600.00 MiB, K (f16): 800.00 MiB, V (f16): 800.00 MiB
[1709429537] llama_new_context_with_model: CPU input buffer size = 15.02 MiB
[1709429537] ggml_backend_metal_buffer_type_alloc_buffer: allocated buffer, size = 204.02 MiB[1709429537] , (14829.75 / 21845.34)[1709429537]
[1709429537] llama_new_context_with_model: Metal compute buffer size = 204.01 MiB
[1709429537] llama_new_context_with_model: CPU compute buffer size = 10.00 MiB
[1709429537] llama_new_context_with_model: graph splits (measure): 3
[1709429537] warming up the model with an empty run