Radek Osmulski radekosmulski

## Alpaca_with_LoRA_bf16.ipynb

      
              1 file
            
          
              0 forks
            
          
              0 comments
            
          
              0 stars
            
          
                radekosmulski
                / Alpaca_with_LoRA_bf16.ipynb
            
            
              Created
              August 25, 2023 21:30
            
          
      View Alpaca_with_LoRA_bf16.ipynb
  
    
      Sorry, something went wrong. Reload?
      Sorry, we cannot display this file.
      Sorry, this file is invalid so it cannot be displayed.
      
          Viewer requires iframe.
      
    
## Alpaca_with_LoRA.ipynb

      
              2 files
            
          
              0 forks
            
          
              0 comments
            
          
              0 stars
            
          
                radekosmulski
                / Alpaca_with_LoRA.ipynb
            
            
              Created
              August 24, 2023 23:34
            
          
      View Alpaca_with_LoRA.ipynb
  
    
      Sorry, something went wrong. Reload?
      Sorry, we cannot display this file.
      Sorry, this file is invalid so it cannot be displayed.
      
          Viewer requires iframe.
      
    
## evaluate_LLM.ipynb

      
              1 file
            
          
              0 forks
            
          
              0 comments
            
          
              0 stars
            
          
                radekosmulski
                / evaluate_LLM.ipynb
            
            
              Created
              July 15, 2023 12:40
            
          
      View evaluate_LLM.ipynb
  
    
      Sorry, something went wrong. Reload?
      Sorry, we cannot display this file.
      Sorry, this file is invalid so it cannot be displayed.
      
          Viewer requires iframe.
      
    
## rmm_mem_mgmt.py
# Initialize RMM pool on ALL workers

def _rmm_pool():
    rmm.reinitialize(
        pool_allocator=True,
        initial_pool_size=None,  # Use default size
    )

client.run(_rmm_pool)

## nvtabular_gpu_cluster.py
# Deploy a Single-Machine Multi-GPU Cluster

protocol = "tcp"  # "tcp" or "ucx"
visible_devices = "0,1,2,3"  # Delect devices to place workers
device_spill_frac = 0.9  # Spill GPU-Worker memory to host at this limit.
# Reduce if spilling fails to prevent
# device memory errors.
cluster = None  # (Optional) Specify existing scheduler port
if cluster is None:
    cluster = LocalCUDACluster(

## Train_a_Two_Tower_model_and_perform_offline_batch_inference.ipynb

      
              1 file
            
          
              2 forks
            
          
              0 comments
            
          
              5 stars
            
          
                radekosmulski
                / Train_a_Two_Tower_model_and_perform_offline_batch_inference.ipynb
            
            
              Last active
              September 8, 2022 17:49
            
          
      View Train_a_Two_Tower_model_and_perform_offline_batch_inference.ipynb
  
    
      Sorry, something went wrong. Reload?
      Sorry, we cannot display this file.
      Sorry, this file is invalid so it cannot be displayed.
      
          Viewer requires iframe.
      
    
## train_and_serve_an_XGBoost_model.ipynb

      
              1 file
            
          
              1 fork
            
          
              0 comments
            
          
              6 stars
            
          
                radekosmulski
                / train_and_serve_an_XGBoost_model.ipynb
            
            
              Last active
              August 29, 2022 04:38
            
          
      View train_and_serve_an_XGBoost_model.ipynb
  
    
      Sorry, something went wrong. Reload?
      Sorry, we cannot display this file.
      Sorry, this file is invalid so it cannot be displayed.
      
          Viewer requires iframe.
      
    
## cpu_vs_gpu_inference.ipynb

      
              1 file
            
          
              0 forks
            
          
              0 comments
            
          
              0 stars
            
          
                radekosmulski
                / cpu_vs_gpu_inference.ipynb
            
            
              Created
              July 22, 2022 03:02
            
          
      View cpu_vs_gpu_inference.ipynb
  
    
      Sorry, something went wrong. Reload?
      Sorry, we cannot display this file.
      Sorry, this file is invalid so it cannot be displayed.
      
          Viewer requires iframe.
      
    
## XGBClassifier_add_noise.ipynb

      
              1 file
            
          
              1 fork
            
          
              0 comments
            
          
              0 stars
            
          
                radekosmulski
                / XGBClassifier_add_noise.ipynb
            
            
              Last active
              June 22, 2022 06:38
            
          
      View XGBClassifier_add_noise.ipynb
  
    
      Sorry, something went wrong. Reload?
      Sorry, we cannot display this file.
      Sorry, this file is invalid so it cannot be displayed.
      
          Viewer requires iframe.
      
    
## visualize_distribution.ipynb

      
              1 file
            
          
              0 forks
            
          
              0 comments
            
          
              0 stars
            
          
                radekosmulski
                / visualize_distribution.ipynb
            
            
              Created
              June 17, 2022 02:31
            
          
      View visualize_distribution.ipynb
  
    
      Sorry, something went wrong. Reload?
      Sorry, we cannot display this file.
      Sorry, this file is invalid so it cannot be displayed.
      
          Viewer requires iframe.
	# Initialize RMM pool on ALL workers

	def _rmm_pool():
	rmm.reinitialize(
	pool_allocator=True,
	initial_pool_size=None, # Use default size
	)

	client.run(_rmm_pool)
	# Deploy a Single-Machine Multi-GPU Cluster

	protocol = "tcp" # "tcp" or "ucx"
	visible_devices = "0,1,2,3" # Delect devices to place workers
	device_spill_frac = 0.9 # Spill GPU-Worker memory to host at this limit.
	# Reduce if spilling fails to prevent
	# device memory errors.
	cluster = None # (Optional) Specify existing scheduler port
	if cluster is None:
	cluster = LocalCUDACluster(