llama-cpp-python not using GPU on colab

[amida47]

Prerequisites

Please answer the following questions for yourself before submitting an issue.

• [ ] I am running the latest code. Development is very rapid so there are no tagged versions as of now.
• [x] I carefully followed the README.md.
• [x] I searched using keywords relevant to my issue to make sure that I am creating a new issue that is not already open (or closed).
• [x] I reviewed the Discussions, and have a new bug or useful enhancement to share.

Expected Behavior

I installed llama-cpp-python[server] using:

• pip install llama-cpp-python[server] \ --extra-index-url https://abetlen.github.io/llama-cpp-python/whl/cu122

I used cu122 after running:

nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2023 NVIDIA Corporation
Built on Tue_Aug_15_22:02:13_PDT_2023
Cuda compilation tools, release 12.2, V12.2.140
Build cuda_12.2.r12.2/compiler.33191640_0

and when I run the server using this command:

• !python3 -m llama_cpp.server --hf_model_repo_id Qwen/Qwen2-7B-Instruct-GGUF --model 'qwen2-7b-instruct-q6_k.gguf' \ --n_ctx 32000 --host 0.0.0.0 --port 8188 --flash_attn True --n_gpu_layers -1

using the --n_gpu_layers -1 should've loaded the model on the gpu

Current Behavior

the model is instead loaded on cpu

llama_model_loader: loaded meta data with 26 key-value pairs and 339 tensors from /root/.cache/huggingface/hub/models--Qwen--Qwen2-7B-Instruct-GGUF/snapshots/ddfd0ef0d00b055363c0fbab0efc1eb2b12186e0/./qwen2-7b-instruct-q6_k.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.name str              = qwen2-7b-instruct
llama_model_loader: - kv   2:                          qwen2.block_count u32              = 28
llama_model_loader: - kv   3:                       qwen2.context_length u32              = 32768
llama_model_loader: - kv   4:                     qwen2.embedding_length u32              = 3584
llama_model_loader: - kv   5:                  qwen2.feed_forward_length u32              = 18944
llama_model_loader: - kv   6:                 qwen2.attention.head_count u32              = 28
llama_model_loader: - kv   7:              qwen2.attention.head_count_kv u32              = 4
llama_model_loader: - kv   8:                       qwen2.rope.freq_base f32              = 1000000.000000
llama_model_loader: - kv   9:     qwen2.attention.layer_norm_rms_epsilon f32              = 0.000001
llama_model_loader: - kv  10:                          general.file_type u32              = 18
llama_model_loader: - kv  11:                       tokenizer.ggml.model str              = gpt2
llama_model_loader: - kv  12:                         tokenizer.ggml.pre str              = qwen2
llama_model_loader: - kv  13:                      tokenizer.ggml.tokens arr[str,152064]  = ["!", "\"", "#", "$", "%", "&", "'", ...
llama_model_loader: - kv  14:                  tokenizer.ggml.token_type arr[i32,152064]  = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv  15:                      tokenizer.ggml.merges arr[str,151387]  = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",...
llama_model_loader: - kv  16:                tokenizer.ggml.eos_token_id u32              = 151645
llama_model_loader: - kv  17:            tokenizer.ggml.padding_token_id u32              = 151643
llama_model_loader: - kv  18:                tokenizer.ggml.bos_token_id u32              = 151643
llama_model_loader: - kv  19:                    tokenizer.chat_template str              = {% for message in messages %}{% if lo...
llama_model_loader: - kv  20:               tokenizer.ggml.add_bos_token bool             = false
llama_model_loader: - kv  21:               general.quantization_version u32              = 2
llama_model_loader: - kv  22:                      quantize.imatrix.file str              = ../Qwen2/gguf/qwen2-7b-imatrix/imatri...
llama_model_loader: - kv  23:                   quantize.imatrix.dataset str              = ../sft_2406.txt
llama_model_loader: - kv  24:             quantize.imatrix.entries_count i32              = 196
llama_model_loader: - kv  25:              quantize.imatrix.chunks_count i32              = 1937
llama_model_loader: - type  f32:  141 tensors
llama_model_loader: - type q6_K:  198 tensors
llm_load_vocab: special tokens cache size = 421
llm_load_vocab: token to piece cache size = 0.9352 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: n_ctx_train      = 32768
llm_load_print_meta: n_embd           = 3584
llm_load_print_meta: n_head           = 28
llm_load_print_meta: n_head_kv        = 4
llm_load_print_meta: n_layer          = 28
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            = 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  = 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       = ?B
llm_load_print_meta: model ftype      = Q6_K
llm_load_print_meta: model params     = 7.62 B
llm_load_print_meta: model size       = 5.82 GiB (6.56 BPW) 
llm_load_print_meta: general.name     = qwen2-7b-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_tensors: ggml ctx size =    0.16 MiB
llm_load_tensors:        CPU buffer size =  5958.79 MiB
warning: failed to mlock 453021696-byte buffer (after previously locking 0 bytes): Cannot allocate memory
Try increasing RLIMIT_MEMLOCK ('ulimit -l' as root).
........................................................................................
llama_new_context_with_model: n_ctx      = 32000
llama_new_context_with_model: n_batch    = 512
llama_new_context_with_model: n_ubatch   = 512
llama_new_context_with_model: flash_attn = 1
llama_new_context_with_model: freq_base  = 1000000.0
llama_new_context_with_model: freq_scale = 1
llama_kv_cache_init:        CPU KV buffer size =  1750.00 MiB
llama_new_context_with_model: KV self size  = 1750.00 MiB, K (f16):  875.00 MiB, V (f16):  875.00 MiB
llama_new_context_with_model:        CPU  output buffer size =     0.59 MiB
llama_new_context_with_model:        CPU compute buffer size =   304.00 MiB
llama_new_context_with_model: graph nodes  = 875
llama_new_context_with_model: graph splits = 1
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 | WASM_SIMD = 0 | BLAS = 0 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 1 | 
Model metadata: {'quantize.imatrix.entries_count': '196', 'quantize.imatrix.dataset': '../sft_2406.txt', 'quantize.imatrix.chunks_count': '1937', 'quantize.imatrix.file': '../Qwen2/gguf/qwen2-7b-imatrix/imatrix.dat', 'tokenizer.ggml.add_bos_token': 'false', 'tokenizer.ggml.bos_token_id': '151643', 'general.architecture': 'qwen2', 'qwen2.block_count': '28', 'qwen2.context_length': '32768', 'tokenizer.chat_template': "{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n' }}{% endif %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}", 'qwen2.attention.head_count_kv': '4', 'tokenizer.ggml.padding_token_id': '151643', 'qwen2.embedding_length': '3584', 'qwen2.attention.layer_norm_rms_epsilon': '0.000001', 'qwen2.attention.head_count': '28', 'tokenizer.ggml.eos_token_id': '151645', 'qwen2.rope.freq_base': '1000000.000000', 'general.file_type': '18', 'general.quantization_version': '2', 'qwen2.feed_forward_length': '18944', 'tokenizer.ggml.model': 'gpt2', 'general.name': 'qwen2-7b-instruct', 'tokenizer.ggml.pre': 'qwen2'}
Available chat formats from metadata: chat_template.default
Using gguf chat template: {% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|im_start|>system
You are a helpful assistant.<|im_end|>
' }}{% endif %}{{'<|im_start|>' + message['role'] + '
' + message['content'] + '<|im_end|>' + '
'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant
' }}{% endif %}
Using chat eos_token: <|im_end|>
Using chat bos_token: <|endoftext|>
INFO:     Started server process [11741]
INFO:     Waiting for application startup.
INFO:     Application startup complete.
INFO:     Uvicorn running on http://0.0.0.0:8188/ (Press CTRL+C to quit)

Environment and Context

Please provide detailed information about your computer setup. This is important in case the issue is not reproducible except for under certain specific conditions.

it is a colab environment with a T4 gpu

update

installing llama-cpp-python using: !CMAKE_ARGS="-DLLAMA_CUDA=on" pip install llama-cpp-python[server]

fixed the problem, but the problem is that it takes 18 mins to install, so using a prebuilt is still preferred, then I am not closing this issue for time being.

[dreambottle]

I think the issue is that there is currently no cuda prebuild of the latest 0.2.78 version, and pip pulls latest by default. I had the same problem installing it on a local machine.

It can be worked around by specifically installing the previous version

pip install --no-cache-dir llama-cpp-python==0.2.77 --extra-index-url https://abetlen.github.io/llama-cpp-python/whl/cu124

Hopefully, the author keeps providing updated cuda builds though.