PygmalionAI
[Bug]: Docker instance doesn't download model (affects VLLM as well)
Your current environment
The output of `python env.py`
```text Collecting environment information... PyTorch version: 2.4.0+cu121 Is debug build: False CUDA used to build PyTorch: 12.1 ROCM used to build PyTorch: N/AOS: Ubuntu 22.04.4 LTS (x86_64) GCC version: (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 Clang version: Could not collect CMake version: Could not collect Libc version: glibc-2.35
Python version: 3.10.12 (main, Nov 6 2024, 20:22:13) [GCC 11.4.0] (64-bit runtime) Python platform: Linux-6.8.0-45-generic-x86_64-with-glibc2.35 Is CUDA available: True CUDA runtime version: Could not collect CUDA_MODULE_LOADING set to: LAZY GPU models and configuration: GPU 0: NVIDIA RTX 6000 Ada Generation GPU 1: NVIDIA RTX 6000 Ada Generation
Nvidia driver version: 550.127.05 cuDNN version: Could not collect HIP runtime version: N/A MIOpen runtime version: N/A Is XNNPACK available: True
CPU: Architecture: x86_64 CPU op-mode(s): 32-bit, 64-bit Address sizes: 48 bits physical, 48 bits virtual Byte Order: Little Endian CPU(s): 128 On-line CPU(s) list: 0-127 Vendor ID: AuthenticAMD Model name: AMD EPYC 75F3 32-Core Processor CPU family: 25 Model: 1 Thread(s) per core: 2 Core(s) per socket: 32 Socket(s): 2 Stepping: 1 Frequency boost: enabled CPU max MHz: 4041.8450 CPU min MHz: 1500.0000 BogoMIPS: 5900.20 Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 pcid sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb cat_l3 cdp_l3 hw_pstate ssbd mba ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 invpcid cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local user_shstk clzero irperf xsaveerptrrdpru wbnoinvd amd_ppin brs arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold v_vmsave_vmload vgif v_spec_ctrl umip pku ospke vaes vpclmulqdq rdpid overflow_recov succor smca debug_swap Virtualization: AMD-V L1d cache: 2 MiB (64 instances) L1i cache: 2 MiB (64 instances) L2 cache: 32 MiB (64 instances) L3 cache: 512 MiB (16 instances) NUMA node(s): 2 NUMA node0 CPU(s): 0-31,64-95 NUMA node1 CPU(s): 32-63,96-127 Vulnerability Gather data sampling: Not affected Vulnerability Itlb multihit: Not affected Vulnerability L1tf: Not affected Vulnerability Mds: Not affected Vulnerability Meltdown: Not affected Vulnerability Mmio stale data: Not affected Vulnerability Reg file data sampling: Not affected Vulnerability Retbleed: Not affected Vulnerability Spec rstack overflow: Mitigation; Safe RET Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl Vulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization Vulnerability Spectre v2: Mitigation; Retpolines; IBPB conditional; IBRS_FW; STIBP always-on; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected Vulnerability Srbds: Not affected Vulnerability Tsx async abort: Not affected
Versions of relevant libraries: [pip3] flashinfer==0.1.4+cu121torch2.4 [pip3] numpy==1.26.4 [pip3] nvidia-nccl-cu12==2.20.5 [pip3] pyzmq==26.2.0 [pip3] torch==2.4.0 [pip3] torchvision==0.19.0 [pip3] transformers==4.45.2 [pip3] triton==3.0.0 [conda] Could not collect ROCM Version: Could not collect Neuron SDK Version: N/A Aphrodite Version: 0.6.4.post1 Aphrodite Build Flags: CUDA Archs: Not Set; ROCm: Disabled; Neuron: Disabled GPU Topology: GPU0 GPU1 NIC0 NIC1 CPU Affinity NUMA Affinity GPU NUMA ID GPU0 X SYS SYS SYS 0-31,64-95 0 N/A GPU1 SYS X NODE NODE 32-63,96-127 1 N/A NIC0 SYS NODE X PIX NIC1 SYS NODE PIX X
Legend:
X = Self SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) PIX = Connection traversing at most a single PCIe bridge NV# = Connection traversing a bonded set of # NVLinks
NIC Legend:
NIC0: mlx5_0 NIC1: mlx5_1```
🐛 Describe the bug
When running alpindale/aphrodite-openai:latest
With arguments
--host 0.0.0.0 --port 8000 --model hf-100/Mistral-Small-Instruct-2409-Spellbound-StoryWriter-22B-instruct-0.4-chkpt-336-16bit --dtype bfloat16 --enforce-eager --gpu-memory-utilization 0.95 --api-key sk-IrR7Bwxtin0haWagUnPrBgq5PurnUz86 --max-model-len 15000 --tensor-parallel-size 2
Aphrodite starts, but hangs before downloading the model
Search
0 matches
2024-12-17T20:43:27.485381977Z INFO: Multiprocessing frontend to use
2024-12-17T20:43:27.485439767Z ipc:///tmp/a931a697-3cfb-48ac-875a-63b02a26afbf for RPC Path.
2024-12-17T20:43:27.488903547Z INFO: Started engine process with PID 85
2024-12-17T20:43:32.099029075Z INFO: Defaulting to use mp for distributed inference.
2024-12-17T20:43:32.102208714Z INFO:
2024-12-17T20:43:32.102225904Z --------------------------------------------------------------------------------
2024-12-17T20:43:32.102229674Z -----
2024-12-17T20:43:32.102598835Z INFO: Initializing Aphrodite Engine (v0.6.4.post1 commit 20f11fd0) with the
2024-12-17T20:43:32.102635205Z following config:
2024-12-17T20:43:32.102960356Z INFO: Model =
2024-12-17T20:43:32.102966976Z 'hf-100/Mistral-Small-Instruct-2409-Spellbound-StoryWriter-22B-instruct-0.4-chkp
2024-12-17T20:43:32.102970136Z t-336-16bit'
2024-12-17T20:43:32.103224377Z INFO: DataType = torch.bfloat16
2024-12-17T20:43:32.103477528Z INFO: Tensor Parallel Size = 2
2024-12-17T20:43:32.103706139Z INFO: Pipeline Parallel Size = 1
2024-12-17T20:43:32.103969639Z INFO: Disable Custom All-Reduce = False
2024-12-17T20:43:32.104205710Z INFO: Context Length = 15000
2024-12-17T20:43:32.104457691Z INFO: Enforce Eager Mode = True
2024-12-17T20:43:32.104693291Z INFO: Prefix Caching = False
2024-12-17T20:43:32.104961542Z INFO: Device = device(type='cuda')
2024-12-17T20:43:32.105294073Z INFO: Guided Decoding Backend =
2024-12-17T20:43:32.105297183Z DecodingConfig(guided_decoding_backend='lm-format-enforcer')
2024-12-17T20:43:32.105577084Z INFO:
2024-12-17T20:43:32.105580624Z --------------------------------------------------------------------------------
2024-12-17T20:43:32.105583044Z -----
2024-12-17T20:43:32.445863267Z WARNING: Reducing Torch parallelism from 64 threads to 1 to avoid unnecessary
2024-12-17T20:43:32.445906567Z CPU contention. Set OMP_NUM_THREADS in the external environment to tune this
2024-12-17T20:43:32.445913698Z value as needed.
2024-12-17T20:43:33.036721517Z (AphroditeWorkerProcess pid=214) INFO: Worker ready; awaiting tasks
I've noticed this happen with vLLM at times too, not sure if it could be related. Unfortunately I haven't found any real pattern to it either
Also running on Runpod.io here, will continue to investigate if I can find a common factor, sorry for the very open ended issue right now
Edit: Also to clarify the vLLM mention, I actually swapped out the docker instance for vLLM, with no settings changes, and got a similar hang with no errors, but the model not downloading
Usually when this happens, restarting the instance will not fix it, but creating a new instance does fix it