问题是什么？
qwen2:72b-instruct-q4_K_M产生垃圾输出：
>>> hello.
#:<,H=&*1(E.E*>G*:^C其他模型在其他量化中工作正常。
Ollama服务器输出：
$ ./ollama serve
2024/07/08 10:07:57 routes.go:1064: INFO server config env="map[CUDA_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: OLLAMA_DEBUG:false OLLAMA_FLASH_ATTENTION:false OLLAMA_HOST:http://127.0.0.1:11434 OLLAMA_INTEL_GPU:false OLLAMA_KEEP_ALIVE: OLLAMA_LLM_LIBRARY: OLLAMA_MAX_LOADED_MODELS:1 OLLAMA_MAX_QUEUE:512 OLLAMA_MAX_VRAM:0 OLLAMA_MODELS:/home/test/.ollama/models OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:1 OLLAMA_ORIGINS:[http://localhost https://localhost http://localhost:* https://localhost:* http://127.0.0.1 https://127.0.0.1 http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http://0.0.0.0:* https://0.0.0.0:* app://* file://* tauri://*] OLLAMA_RUNNERS_DIR: OLLAMA_SCHED_SPREAD:false OLLAMA_TMPDIR: ROCR_VISIBLE_DEVICES:]"
time=2024-07-08T10:07:57.658+02:00 level=INFO source=images.go:730 msg="total blobs: 30"
time=2024-07-08T10:07:57.661+02:00 level=INFO source=images.go:737 msg="total unused blobs removed: 0"
time=2024-07-08T10:07:57.661+02:00 level=INFO source=routes.go:1111 msg="Listening on 127.0.0.1:11434 (version 0.1.48)"
time=2024-07-08T10:07:57.662+02:00 level=INFO source=payload.go:30 msg="extracting embedded files" dir=/tmp/ollama2929279884/runners
time=2024-07-08T10:08:00.530+02:00 level=INFO source=payload.go:44 msg="Dynamic LLM libraries [cpu cpu_avx cpu_avx2 cuda_v11 rocm_v60101]"
time=2024-07-08T10:08:00.679+02:00 level=INFO source=types.go:98 msg="inference compute" id=GPU-7b971568-6cd2-b804-d8eb-902eb8689068 library=cuda compute=6.1 driver=12.2 name="NVIDIA GeForce GTX 1080 Ti" total="10.9 GiB" available="10.4 GiB"
[GIN] 2024/07/08 - 10:08:22 | 200 | 81.665µs | 127.0.0.1 | HEAD "/"
[GIN] 2024/07/08 - 10:08:22 | 200 | 37.454138ms | 127.0.0.1 | POST "/api/show"
time=2024-07-08T10:08:22.768+02:00 level=INFO source=memory.go:309 msg="offload to cuda" layers.requested=-1 layers.model=81 layers.offload=14 layers.split="" memory.available="[10.4 GiB]" memory.required.full="49.6 GiB" memory.required.partial="10.2 GiB" memory.required.kv="640.0 MiB" memory.required.allocations="[10.2 GiB]" memory.weights.total="43.2 GiB" memory.weights.repeating="42.2 GiB" memory.weights.nonrepeating="974.6 MiB" memory.graph.full="313.0 MiB" memory.graph.partial="1.3 GiB"
time=2024-07-08T10:08:22.769+02:00 level=INFO source=server.go:368 msg="starting llama server" cmd="/tmp/ollama2929279884/runners/cuda_v11/ollama_llama_server --model /home/test/.ollama/models/blobs/sha256-59e062dadfebe1e1b7dae3aa2ed6f60190c03e9738451e6963d74a5aa6a464a9 --ctx-size 2048 --batch-size 512 --embedding --log-disable --n-gpu-layers 14 --no-mmap --parallel 1 --port 45747"
time=2024-07-08T10:08:22.769+02:00 level=INFO source=sched.go:382 msg="loaded runners" count=1
time=2024-07-08T10:08:22.769+02:00 level=INFO source=server.go:556 msg="waiting for llama runner to start responding"
time=2024-07-08T10:08:22.770+02:00 level=INFO source=server.go:594 msg="waiting for server to become available" status="llm server error"
INFO [main] build info | build=1 commit="7c26775" tid="139967113396224" timestamp=1720426102
INFO [main] system info | n_threads=8 n_threads_batch=-1 system_info="AVX = 1 | AVX_VNNI = 0 | AVX2 = 0 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | AVX512_BF16 = 0 | FMA = 0 | NEON = 0 | SVE = 0 | ARM_FMA = 0 | F16C = 0 | FP16_VA = 0 | WASM_SIMD = 0 | BLAS = 1 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 1 | " tid="139967113396224" timestamp=1720426102 total_threads=16
INFO [main] HTTP server listening | hostname="127.0.0.1" n_threads_http="15" port="45747" tid="139967113396224" timestamp=1720426102
llama_model_loader: loaded meta data with 21 key-value pairs and 963 tensors from /home/test/.ollama/models/blobs/sha256-59e062dadfebe1e1b7dae3aa2ed6f60190c03e9738451e6963d74a5aa6a464a9 (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 = Qwen2-72B-Instruct
llama_model_loader: - kv 2: qwen2.block_count u32 = 80
llama_model_loader: - kv 3: qwen2.context_length u32 = 32768
llama_model_loader: - kv 4: qwen2.embedding_length u32 = 8192
llama_model_loader: - kv 5: qwen2.feed_forward_length u32 = 29568
llama_model_loader: - kv 6: qwen2.attention.head_count u32 = 64
llama_model_loader: - kv 7: qwen2.attention.head_count_kv u32 = 8
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 = 15
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: general.quantization_version u32 = 2
llama_model_loader: - type f32: 401 tensors
llama_model_loader: - type q5_0: 40 tensors
llama_model_loader: - type q8_0: 40 tensors
llama_model_loader: - type q4_K: 401 tensors
llama_model_loader: - type q5_K: 40 tensors
llama_model_loader: - type q6_K: 41 tensors
time=2024-07-08T10:08:23.021+02:00 level=INFO source=server.go:594 msg="waiting for server to become available" status="llm server loading model"
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 = 8192
llm_load_print_meta: n_head = 64
llm_load_print_meta: n_head_kv = 8
llm_load_print_meta: n_layer = 80
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 = 8
llm_load_print_meta: n_embd_k_gqa = 1024
llm_load_print_meta: n_embd_v_gqa = 1024
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 = 29568
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 = 70B
llm_load_print_meta: model ftype = Q4_K - Medium
llm_load_print_meta: model params = 72.71 B
llm_load_print_meta: model size = 44.15 GiB (5.22 BPW)
llm_load_print_meta: general.name = Qwen2-72B-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|>'
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: yes
ggml_cuda_init: CUDA_USE_TENSOR_CORES: no
ggml_cuda_init: found 1 CUDA devices:
Device 0: NVIDIA GeForce GTX 1080 Ti, compute capability 6.1, VMM: yes
llm_load_tensors: ggml ctx size = 0.92 MiB
time=2024-07-08T10:08:24.478+02:00 level=INFO source=server.go:594 msg="waiting for server to become available" status="llm server not responding"
time=2024-07-08T10:08:31.046+02:00 level=INFO source=server.go:594 msg="waiting for server to become available" status="llm server loading model"
llm_load_tensors: offloading 14 repeating layers to GPU
llm_load_tensors: offloaded 14/81 layers to GPU
llm_load_tensors: CUDA_Host buffer size = 37150.14 MiB
llm_load_tensors: CUDA0 buffer size = 8063.30 MiB
llama_new_context_with_model: n_ctx = 2048
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 = 1000000.0
llama_new_context_with_model: freq_scale = 1
llama_kv_cache_init: CUDA_Host KV buffer size = 528.00 MiB
llama_kv_cache_init: CUDA0 KV buffer size = 112.00 MiB
llama_new_context_with_model: KV self size = 640.00 MiB, K (f16): 320.00 MiB, V (f16): 320.00 MiB
llama_new_context_with_model: CUDA_Host output buffer size = 0.61 MiB
llama_new_context_with_model: CUDA0 compute buffer size = 1287.53 MiB
llama_new_context_with_model: CUDA_Host compute buffer size = 20.01 MiB
llama_new_context_with_model: graph nodes = 2806
llama_new_context_with_model: graph splits = 928
INFO [main] model loaded | tid="139967113396224" timestamp=1720426189
time=2024-07-08T10:09:49.209+02:00 level=INFO source=server.go:599 msg="llama runner started in 86.44 seconds"
[GIN] 2024/07/08 - 10:09:49 | 200 | 1m26s | 127.0.0.1 | POST "/api/chat"
[GIN] 2024/07/08 - 10:11:28 | 200 | 47.974392692s | 127.0.0.1 | POST "/api/chat"
[GIN] 2024/07/08 - 10:12:52 | 200 | 20.454650534s | 127.0.0.1 | POST "/api/chat"操作系统
Linux
GPU
Nvidia
CPU
Intel
Ollama版本
0.1.48
5条答案
按热度按时间ui7jx7zq1#
PS: 有趣。我尝试通过在调试模式下运行ollama服务器来提供更多上下文，并且在后续运行中无法重现垃圾输出。似乎存在一些不稳定问题。
q3qa4bjr2#
PPS:似乎提示需要更长一些，以使这个问题更有可能出现。这是调试输出：
mkshixfv3#
相同的模型，不同的型号。
quhf5bfb4#
也正在经历它 #5641 (评论),使用下面的上下文，但大约为qwen2的> 10k。
a0x5cqrl5#
经过一段时间的正常工作，然后突然开始产生乱码。在使用codegeex和qwen 2时遇到了这个问题。