Upload folder using huggingface_hub (#2)

- cad3df85b7b2334b2857c03565330c39ce1c874a0d2ad8241e264f1475f925f1 (5278bc78862e43f0b2524b7fc61aeb8302f5916a)

README.md

@@ -59,16 +59,16 @@ You can run the smashed model with these steps:
     ```
 2. Load & run the model.
     ```python 
-  from transformers import AutoModelForCausalLM, AutoTokenizer
-  from awq import AutoAWQForCausalLM
-
-  model = AutoAWQForCausalLM.from_quantized("PrunaAI/OpenLLM-France-Lucie-7B-AWQ-4bit-smashed", trust_remote_code=True, device_map='auto')
-  tokenizer = AutoTokenizer.from_pretrained("OpenLLM-France/Lucie-7B")
-
-  input_ids = tokenizer("Quelle est la couleur des pruneaux?", return_tensors='pt').to(next(model.parameters()).device)["input_ids"]
-
-  outputs = model.generate(input_ids, max_new_tokens=216)
-  tokenizer.decode(outputs[0])
+   from transformers import AutoModelForCausalLM, AutoTokenizer
+   from awq import AutoAWQForCausalLM
+
+   model = AutoAWQForCausalLM.from_quantized("PrunaAI/OpenLLM-France-Lucie-7B-Instruct-AWQ-4bit-smashed", trust_remote_code=True, device_map='auto')
+   tokenizer = AutoTokenizer.from_pretrained("OpenLLM-France/Lucie-7B-Instruct")
+    
+   input_ids = tokenizer("What is the color of prunes?,", return_tensors='pt').to(model.device)["input_ids"]
+    
+   outputs = model.generate(input_ids, max_new_tokens=216)
+   tokenizer.decode(outputs[0])
     ```
 
 ## Configurations

base_results.json

@@ -0,0 +1,18 @@
+{
+    "current_gpu_type": "NVIDIA L40S",
+    "current_gpu_total_memory": 45372.6875,
+    "memory_inference_first": 25590.0,
+    "memory_inference": 25590.0,
+    "token_generation_latency_sync": 30.846586608886717,
+    "token_generation_latency_async": 30.805427208542824,
+    "token_generation_throughput_sync": 0.03241849779618423,
+    "token_generation_throughput_async": 0.03246181243422862,
+    "token_generation_CO2_emissions": 1.7413158175960427e-06,
+    "token_generation_energy_consumption": 0.0012653706804628014,
+    "inference_latency_sync": 25.427967834472657,
+    "inference_latency_async": 24.5699405670166,
+    "inference_throughput_sync": 0.039326776190282166,
+    "inference_throughput_async": 0.04070013915061842,
+    "inference_CO2_emissions": 2.974451769021189e-06,
+    "inference_energy_consumption": 1.0356497508141638e-05
+}

smashed_results.json

@@ -0,0 +1,18 @@
+{
+    "current_gpu_type": "NVIDIA L40S",
+    "current_gpu_total_memory": 45372.6875,
+    "memory_inference_first": 4312.0,
+    "memory_inference": 4208.0,
+    "token_generation_latency_sync": 47.82877426147461,
+    "token_generation_latency_async": 48.09337295591831,
+    "token_generation_throughput_sync": 0.020907916112027268,
+    "token_generation_throughput_async": 0.02079288555861086,
+    "token_generation_CO2_emissions": 3.164995601682604e-06,
+    "token_generation_energy_consumption": 0.00230724002830658,
+    "inference_latency_sync": 41.39498291015625,
+    "inference_latency_async": 40.357160568237305,
+    "inference_throughput_sync": 0.02415751691866625,
+    "inference_throughput_async": 0.024778750187569934,
+    "inference_CO2_emissions": 3.351062981628799e-06,
+    "inference_energy_consumption": 1.1458303099082523e-05
+}