deepseek-ai_DeepSeek-V3_86518964eaef84e3fdd98e9861759a1384f9c29d

deepseek-ai_DeepSeek-V3_86518964eaef84e3fdd98e9861759a1384f9c29d.logits.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1993feb3897078d1269140657755c892fdf86e68cbc040c4997a4b73b78524f8
+size 437566024

run_test.py

@@ -1,43 +1,51 @@
 #!/usr/bin/env python3
 import os, sys
-import transformers, torch, tqdm
+import accelerate, safetensors.torch, transformers, torch, tqdm
 
 model_id, revision = sys.argv[1:]
 user, model = model_id.split('/')
 
-dir = f'{user}_{model}_{revision}'
-os.makedirs(dir, exist_ok=True)
-os.chdir(dir)
+fn = f'{user}_{model}_{revision}.logits.safetensors'
 
-pipe = transformers.pipeline('text-generation', model_id, revision=revision, trust_remote_code=True, device_map='auto', torch_dtype='auto')
+config = transformers.AutoConfig.from_pretrained(model_id, revision=revision, trust_remote_code=True)
+tokenizer = transformers.AutoTokenizer.from_pretrained(model_id, revision=revision, trust_remote_code=True)
+Model = transformers.AutoModelForCausalLM
+if config.model_type == 'deepseek_v3':
+    #Model = transformers.DeepseekV3ForCausalLM
+    pass
+model = Model.from_pretrained(model_id, trust_remote_code=True, torch_dtype='auto', device_map='cpu')
+if config.model_type == 'deepseek_v3':
+    model._supports_cache_class = False
+model = accelerate.cpu_offload(model, 'cuda:0', offload_buffers=True)
 
+pipe = transformers.pipeline('text-generation', model=model, config=config, tokenizer=tokenizer)
+
+tensors = {}
 def store_tensor(descr, tensor):
-    shapestr = 'x'.join([str(i) for i in tensor.shape])
-    dtypestr = str(tensor.dtype).rsplit('.',1)[-1]
-    fn = f'{descr}.{shapestr}.{dtypestr}.mmap'
-    mmap = torch.from_file(fn, shared=True, dtype=tensor.dtype, size=tensor.numel())
-    mmap[:] = tensor.flatten()
-    print(fn, end='\r', flush=True)
+    tensors[descr] = tensor.cpu().detach().contiguous()
 
 IDX = 0
 module_names = {mod:name for name, mod in pipe.model.named_modules()}
+tensors = {}
 def hook(module, inputs, outputs):
     global IDX
     name = module_names[module]
     for idx, input in enumerate(inputs):
         if isinstance(input, torch.Tensor):
-            store_tensor(f'{IDX:07d}.{name}.input.{idx}', input);
+            store_tensor(f'{name}.input.{idx}', input);
     if isinstance(outputs, torch.Tensor):
-        store_tensor(f'{IDX:07d}.{name}.output', outputs);
+        store_tensor(f'{name}.output', outputs);
     else:
         for idx, output in enumerate(outputs):
             if isinstance(output, torch.Tensor):
-                store_tensor(f'{IDX:07d}.{name}.output.{idx}', output);
+                store_tensor(f'{name}.output.{idx}', output);
     IDX += 1
 
 for module in pipe.model.modules():
     module.register_forward_hook(hook)
 
-output = pipe('Once upon a time,', do_sample=False, max_new_tokens=1)
+prompt = 'Once upon a time,'
+output = pipe(prompt, do_sample=False, max_new_tokens=1, temperature=1.0, top_p=1.0)
+safetensors.torch.save_file(tensors, fn, dict(model_id=model_id, revision=revision, prompt=prompt))
 print()
 print(output)