Hugging Face's logo

Spaces:
yusufs
/
llama32-3b-instruct
Paused

App Files Community
yusufs commited on
Commit
852d95b
·
verified ·
1 Parent(s): 637a86f

Update Dockerfile

Browse files
Files changed (1) hide show
  1. Dockerfile +116 -1
Dockerfile CHANGED
@@ -61,7 +61,122 @@ RUN pip install uv setuptools
61
 
62
  # Install vLLM
63
  # RUN uv pip install --system vllm==0.10.0 torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu113
64
- RUN uv pip install --system --index-strategy unsafe-best-match vllm==0.10.0 --extra-index-url https://download.pytorch.org/whl/cu128
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
65
 
66
  # # Then, install xformers with the --no-build-isolation flag
67
  # RUN uv pip install --system \
 
61
 
62
  # Install vLLM
63
  # RUN uv pip install --system vllm==0.10.0 torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu113
64
+ # Downgrade triton because following error occured when using triton==3.3.1
65
+ # /usr/local/lib/python3.12/dist-packages/vllm/attention/ops/prefix_prefill.py:36:0: error: Failures have been detected while processing an MLIR pass pipeline
66
+ # /usr/local/lib/python3.12/dist-packages/vllm/attention/ops/prefix_prefill.py:36:0: note: Pipeline failed while executing [`ConvertTritonGPUToLLVM` on 'builtin.module' operation]: reproducer generated at `std::errs, please share the reproducer above with Triton project.`
67
+ # INFO: 10.16.9.222:28100 - "POST /v1/chat/completions HTTP/1.1" 500 Internal Server Error
68
+ # ERROR 08-06 19:13:13 [engine.py:165] RuntimeError('PassManager::run failed')
69
+ # ERROR 08-06 19:13:13 [engine.py:165] Traceback (most recent call last):
70
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/vllm/engine/multiprocessing/engine.py", line 163, in start
71
+ # ERROR 08-06 19:13:13 [engine.py:165] self.run_engine_loop()
72
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/vllm/engine/multiprocessing/engine.py", line 226, in run_engine_loop
73
+ # ERROR 08-06 19:13:13 [engine.py:165] request_outputs = self.engine_step()
74
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^
75
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/vllm/engine/multiprocessing/engine.py", line 252, in engine_step
76
+ # ERROR 08-06 19:13:13 [engine.py:165] raise e
77
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/vllm/engine/multiprocessing/engine.py", line 235, in engine_step
78
+ # ERROR 08-06 19:13:13 [engine.py:165] return self.engine.step()
79
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^
80
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/vllm/engine/llm_engine.py", line 1334, in step
81
+ # ERROR 08-06 19:13:13 [engine.py:165] outputs = self.model_executor.execute_model(
82
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
83
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/vllm/executor/executor_base.py", line 146, in execute_model
84
+ # ERROR 08-06 19:13:13 [engine.py:165] output = self.collective_rpc("execute_model",
85
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
86
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/vllm/executor/uniproc_executor.py", line 58, in collective_rpc
87
+ # ERROR 08-06 19:13:13 [engine.py:165] answer = run_method(self.driver_worker, method, args, kwargs)
88
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
89
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/vllm/utils/__init__.py", line 2985, in run_method
90
+ # ERROR 08-06 19:13:13 [engine.py:165] return func(*args, **kwargs)
91
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^
92
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/vllm/worker/worker_base.py", line 417, in execute_model
93
+ # ERROR 08-06 19:13:13 [engine.py:165] output = self.model_runner.execute_model(
94
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
95
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/torch/utils/_contextlib.py", line 116, in decorate_context
96
+ # ERROR 08-06 19:13:13 [engine.py:165] return func(*args, **kwargs)
97
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^
98
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/vllm/worker/model_runner.py", line 1703, in execute_model
99
+ # ERROR 08-06 19:13:13 [engine.py:165] hidden_or_intermediate_states = model_executable(
100
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^
101
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl
102
+ # ERROR 08-06 19:13:13 [engine.py:165] return self._call_impl(*args, **kwargs)
103
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
104
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/torch/nn/modules/module.py", line 1762, in _call_impl
105
+ # ERROR 08-06 19:13:13 [engine.py:165] return forward_call(*args, **kwargs)
106
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
107
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/vllm/model_executor/models/llama.py", line 584, in forward
108
+ # ERROR 08-06 19:13:13 [engine.py:165] model_output = self.model(input_ids, positions, intermediate_tensors,
109
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
110
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/vllm/compilation/decorators.py", line 206, in __call__
111
+ # ERROR 08-06 19:13:13 [engine.py:165] return self.forward(*args, **kwargs)
112
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
113
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/vllm/model_executor/models/llama.py", line 392, in forward
114
+ # ERROR 08-06 19:13:13 [engine.py:165] hidden_states, residual = layer(positions, hidden_states, residual)
115
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
116
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl
117
+ # ERROR 08-06 19:13:13 [engine.py:165] return self._call_impl(*args, **kwargs)
118
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
119
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/torch/nn/modules/module.py", line 1762, in _call_impl
120
+ # ERROR 08-06 19:13:13 [engine.py:165] return forward_call(*args, **kwargs)
121
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
122
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/vllm/model_executor/models/llama.py", line 305, in forward
123
+ # ERROR 08-06 19:13:13 [engine.py:165] hidden_states = self.self_attn(positions=positions,
124
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
125
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl
126
+ # ERROR 08-06 19:13:13 [engine.py:165] return self._call_impl(*args, **kwargs)
127
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
128
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/torch/nn/modules/module.py", line 1762, in _call_impl
129
+ # ERROR 08-06 19:13:13 [engine.py:165] return forward_call(*args, **kwargs)
130
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
131
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/vllm/model_executor/models/llama.py", line 203, in forward
132
+ # ERROR 08-06 19:13:13 [engine.py:165] attn_output = self.attn(q, k, v)
133
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^
134
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/torch/nn/modules/module.py", line 1751, in _wrapped_call_impl
135
+ # ERROR 08-06 19:13:13 [engine.py:165] return self._call_impl(*args, **kwargs)
136
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
137
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/torch/nn/modules/module.py", line 1762, in _call_impl
138
+ # ERROR 08-06 19:13:13 [engine.py:165] return forward_call(*args, **kwargs)
139
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
140
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/vllm/attention/layer.py", line 288, in forward
141
+ # ERROR 08-06 19:13:13 [engine.py:165] return torch.ops.vllm.unified_attention(
142
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
143
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/torch/_ops.py", line 1158, in __call__
144
+ # ERROR 08-06 19:13:13 [engine.py:165] return self._op(*args, **(kwargs or {}))
145
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
146
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/vllm/attention/layer.py", line 448, in unified_attention
147
+ # ERROR 08-06 19:13:13 [engine.py:165] output = self.impl.forward(self, query, key, value, kv_cache,
148
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
149
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/vllm/attention/backends/xformers.py", line 584, in forward
150
+ # ERROR 08-06 19:13:13 [engine.py:165] out = PagedAttention.forward_prefix(
151
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
152
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/vllm/attention/ops/paged_attn.py", line 214, in forward_prefix
153
+ # ERROR 08-06 19:13:13 [engine.py:165] context_attention_fwd(
154
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/torch/utils/_contextlib.py", line 116, in decorate_context
155
+ # ERROR 08-06 19:13:13 [engine.py:165] return func(*args, **kwargs)
156
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^
157
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/vllm/attention/ops/prefix_prefill.py", line 850, in context_attention_fwd
158
+ # ERROR 08-06 19:13:13 [engine.py:165] _fwd_kernel[grid](
159
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/triton/runtime/jit.py", line 347, in <lambda>
160
+ # ERROR 08-06 19:13:13 [engine.py:165] return lambda *args, **kwargs: self.run(grid=grid, warmup=False, *args, **kwargs)
161
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
162
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/triton/runtime/jit.py", line 569, in run
163
+ # ERROR 08-06 19:13:13 [engine.py:165] kernel = self.compile(src, target=target, options=options.__dict__)
164
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
165
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/triton/compiler/compiler.py", line 284, in compile
166
+ # ERROR 08-06 19:13:13 [engine.py:165] next_module = compile_ir(module, metadata)
167
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
168
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/triton/backends/nvidia/compiler.py", line 450, in <lambda>
169
+ # ERROR 08-06 19:13:13 [engine.py:165] stages["llir"] = lambda src, metadata: self.make_llir(src, metadata, options, capability)
170
+ # ERROR 08-06 19:13:13 [engine.py:165] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
171
+ # ERROR 08-06 19:13:13 [engine.py:165] File "/usr/local/lib/python3.12/dist-packages/triton/backends/nvidia/compiler.py", line 341, in make_llir
172
+ # ERROR 08-06 19:13:13 [engine.py:165] pm.run(mod)
173
+ # ERROR 08-06 19:13:13 [engine.py:165] RuntimeError: PassManager::run failed
174
+ # DEBUG 08-06 19:13:13 [engine.py:169] MQLLMEngine is shut down.
175
+ # INFO: Shutting down
176
+ # INFO: Waiting for application shutdown.
177
+ # INFO: Application shutdown complete.
178
+ # INFO: Finished server process [27]
179
+ RUN uv pip install --system --index-strategy unsafe-best-match vllm==0.10.0 triton==3.2 --extra-index-url https://download.pytorch.org/whl/cu128
180
 
181
  # # Then, install xformers with the --no-build-isolation flag
182
  # RUN uv pip install --system \