search

perk11 / large-model-proxy

Large Model Proxy is designed to make it easy to run multiple resource-heavy Large Models (LM) on the same machine with limited amount of VRAM/other resources. It listens on a dedicated port for each proxied LM, making them always available to the clients connecting to these ports.
GNU General Public License v2.0
46 stars 3 forks source link

No response/timeout from the proxy service #19

Closed auspiciousdiviner closed 1 month ago

auspiciousdiviner commented 1 month ago

The service runs, receives connection, and loads models according to the logs but doesn't do anything at all or respond to any chat messages:

$ /ai/large-model-proxy/large-model-proxy -c /ai/llm-proxy-config.json
2024/10/07 02:22:52 [Qwen 2.5 coder] Listening on port 7860
2024/10/07 02:22:56 [Qwen 2.5 coder] New client connection received 127.0.0.1:7860->127.0.0.1:54132
2024/10/07 02:22:56 [Qwen 2.5 coder] Reserving VRAM-GPU-1: 5000, RAM: 3000
2024/10/07 02:22:56 [Qwen 2.5 coder] Starting "/ai/llama.cpp/llama-server -m /ai/catalog/Qwen2.5/Qwen2.5-Coder-7B-Instruct-Q4_K_M.gguf -c 8192 -ngl 100 --port 17860 --host 0.0.0.0", log file: logs/Qwen 2.5 coder.log, workdir:
2024/10/07 02:22:56 [Qwen 2.5 coder] Running healthcheck command "curl --fail http://localhost:7860/"
2024/10/07 02:22:56 [Qwen 2.5 coder] New client connection received [::1]:7860->[::1]:32798
2024/10/07 02:23:56 [Qwen 2.5 coder] New client connection received 127.0.0.1:7860->127.0.0.1:60008

My config:

{
  "MaxTimeToWaitForServiceToCloseConnectionBeforeGivingUpSeconds": 1200,
  "ShutDownAfterInactivitySeconds": 120,
  "ResourcesAvailable": {
     "VRAM-GPU-1": 16000,
     "RAM": 20000
  },
  "Services": [
    {
      "Name": "Qwen 2.5 coder",
      "ListenPort": "7860",
      "ProxyTargetHost": "localhost",
      "ProxyTargetPort": "17860",
      "Command": "/ai/llama.cpp/llama-server",
      "Args": "-m /ai/catalog/Qwen2.5/Qwen2.5-Coder-7B-Instruct-Q4_K_M.gguf -c 8192 -ngl 100 --port 17860 --host 0.0.0.0",
      "HealthcheckCommand": "curl --fail http://localhost:7860/",
      "HealthcheckIntervalMilliseconds": 200,
      "RestartOnConnectionFailure": false,
      "ResourceRequirements": {
        "VRAM-GPU-1": 5000,
        "RAM": 3000
      }
    }
  ]
}

Qwen 2.5 coder log:

ggml_cuda_init: GGML_CUDA_FORCE_MMQ:    no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
  Device 0: AMD Radeon RX 6800, compute capability 10.3, VMM: no
build: 3870 (841713e1) with cc (GCC) 14.2.1 20240910 for x86_64-pc-linux-gnu
system info: n_threads = 8, n_threads_batch = 8, total_threads = 16

system_info: n_threads = 8 (n_threads_batch = 8) / 16 | AVX = 1 | AVX_VNNI = 0 | AVX2 = 1 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | AVX512_BF16 = 0 | FMA = 1 | NEON = 0 | SVE = 0 | ARM_FMA = 0 | F16C = 1 | FP16_VA = 0 | RISCV_VECT = 0 | WASM_SIMD = 0 | BLAS = 1 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 1 | 

main: HTTP server is listening, hostname: 0.0.0.0, port: 17860, http threads: 15
main: loading model
llama_model_loader: loaded meta data with 30 key-value pairs and 339 tensors from /ai/catalog/Qwen2.5/Qwen2.5-Coder-7B-Instruct-Q4_K_M.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              = qwen2
llama_model_loader: - kv   1:                               general.type str              = model
llama_model_loader: - kv   2:                               general.name str              = Qwen2.5 Coder 7B Instruct
llama_model_loader: - kv   3:                           general.finetune str              = Instruct
llama_model_loader: - kv   4:                           general.basename str              = Qwen2.5-Coder
llama_model_loader: - kv   5:                         general.size_label str              = 7B
llama_model_loader: - kv   6:                          qwen2.block_count u32              = 28
llama_model_loader: - kv   7:                       qwen2.context_length u32              = 131072
llama_model_loader: - kv   8:                     qwen2.embedding_length u32              = 3584
llama_model_loader: - kv   9:                  qwen2.feed_forward_length u32              = 18944
llama_model_loader: - kv  10:                 qwen2.attention.head_count u32              = 28
llama_model_loader: - kv  11:              qwen2.attention.head_count_kv u32              = 4
llama_model_loader: - kv  12:                       qwen2.rope.freq_base f32              = 1000000.000000
llama_model_loader: - kv  13:     qwen2.attention.layer_norm_rms_epsilon f32              = 0.000001
llama_model_loader: - kv  14:                          general.file_type u32              = 15
llama_model_loader: - kv  15:                       tokenizer.ggml.model str              = gpt2
llama_model_loader: - kv  16:                         tokenizer.ggml.pre str              = qwen2
llama_model_loader: - kv  17:                      tokenizer.ggml.tokens arr[str,152064]  = ["!", "\"", "#", "$", "%", "&", "'", ...
llama_model_loader: - kv  18:                  tokenizer.ggml.token_type arr[i32,152064]  = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv  19:                      tokenizer.ggml.merges arr[str,151387]  = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",...
llama_model_loader: - kv  20:                tokenizer.ggml.eos_token_id u32              = 151645
llama_model_loader: - kv  21:            tokenizer.ggml.padding_token_id u32              = 151643
llama_model_loader: - kv  22:                tokenizer.ggml.bos_token_id u32              = 151643
llama_model_loader: - kv  23:               tokenizer.ggml.add_bos_token bool             = false
llama_model_loader: - kv  24:                    tokenizer.chat_template str              = {%- if tools %}\n    {{- '<|im_start|>...
llama_model_loader: - kv  25:               general.quantization_version u32              = 2
llama_model_loader: - kv  26:                      quantize.imatrix.file str              = /models_out/Qwen2.5-Coder-7B-Instruct...
llama_model_loader: - kv  27:                   quantize.imatrix.dataset str              = /training_dir/calibration_datav3.txt
llama_model_loader: - kv  28:             quantize.imatrix.entries_count i32              = 196
llama_model_loader: - kv  29:              quantize.imatrix.chunks_count i32              = 128
llama_model_loader: - type  f32:  141 tensors
llama_model_loader: - type q4_K:  169 tensors
llama_model_loader: - type q6_K:   29 tensors
llm_load_vocab: special tokens cache size = 22
llm_load_vocab: token to piece cache size = 0.9310 MB
llm_load_print_meta: format           = GGUF V3 (latest)
llm_load_print_meta: arch             = qwen2
llm_load_print_meta: vocab type       = BPE
llm_load_print_meta: n_vocab          = 152064
llm_load_print_meta: n_merges         = 151387
llm_load_print_meta: vocab_only       = 0
llm_load_print_meta: n_ctx_train      = 131072
llm_load_print_meta: n_embd           = 3584
llm_load_print_meta: n_layer          = 28
llm_load_print_meta: n_head           = 28
llm_load_print_meta: n_head_kv        = 4
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            = 7
llm_load_print_meta: n_embd_k_gqa     = 512
llm_load_print_meta: n_embd_v_gqa     = 512
llm_load_print_meta: f_norm_eps       = 0.0e+00
llm_load_print_meta: f_norm_rms_eps   = 1.0e-06
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             = 18944
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        = 2
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  = 131072
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      = Q4_K - Medium
llm_load_print_meta: model params     = 7.62 B
llm_load_print_meta: model size       = 4.36 GiB (4.91 BPW) 
llm_load_print_meta: general.name     = Qwen2.5 Coder 7B Instruct
llm_load_print_meta: BOS token        = 151643 '<|endoftext|>'
llm_load_print_meta: EOS token        = 151645 '<|im_end|>'
llm_load_print_meta: PAD token        = 151643 '<|endoftext|>'
llm_load_print_meta: LF token         = 148848 'ÄĬ'
llm_load_print_meta: EOT token        = 151645 '<|im_end|>'
llm_load_print_meta: EOG token        = 151643 '<|endoftext|>'
llm_load_print_meta: EOG token        = 151645 '<|im_end|>'
llm_load_print_meta: max token length = 256
llm_load_tensors: ggml ctx size =    0.30 MiB
llm_load_tensors: offloading 28 repeating layers to GPU
llm_load_tensors: offloading non-repeating layers to GPU
llm_load_tensors: offloaded 29/29 layers to GPU
llm_load_tensors:      ROCm0 buffer size =  4168.09 MiB
llm_load_tensors:        CPU buffer size =   292.36 MiB
..................................................................................
llama_new_context_with_model: n_ctx      = 8192
llama_new_context_with_model: n_batch    = 2048
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:      ROCm0 KV buffer size =   448.00 MiB
llama_new_context_with_model: KV self size  =  448.00 MiB, K (f16):  224.00 MiB, V (f16):  224.00 MiB
llama_new_context_with_model:  ROCm_Host  output buffer size =     1.16 MiB
llama_new_context_with_model:      ROCm0 compute buffer size =   492.00 MiB
llama_new_context_with_model:  ROCm_Host compute buffer size =    23.01 MiB
llama_new_context_with_model: graph nodes  = 986
llama_new_context_with_model: graph splits = 2
llama_init_from_gpt_params: warming up the model with an empty run - please wait ... (--no-warmup to disable)
srv          init: initializing slots, n_slots = 1
slot         init: id  0 | task -1 | new slot n_ctx_slot = 8192
main: model loaded
main: chat template, built_in: 1, chat_example: '<|im_start|>system
You are a helpful assistant<|im_end|>
<|im_start|>user
Hello<|im_end|>
<|im_start|>assistant
Hi there<|im_end|>
<|im_start|>user
How are you?<|im_end|>
<|im_start|>assistant
'
main: server is listening on 0.0.0.0:17860 - starting the main loop
srv  update_slots: all slots are idle
perk11 commented 1 month ago

Thanks for posting, I think I made a mistake in my Readme healthcheck example. Can you try changing the healthcheck to: "HealthcheckCommand": " "curl --fail http://localhost:17860/health",

(note, the port is 17860, not 7860)

auspiciousdiviner commented 1 month ago

Hah, wow, instantly worked after that change. That's crazy.