Training checkpoint 8000 - 0.1B tokens

README.md

@@ -18,11 +18,11 @@ pipeline_tag: text-generation
 
 ## Current Training Status
 
-- **Training Step**: 6,000
-- **Tokens Processed**: 0.06B tokens
-- **Current Loss**: 4.7385
+- **Training Step**: 8,000
+- **Tokens Processed**: 0.08B tokens
+- **Current Loss**: 4.7271
 - **Spike Rate**: 0.0113
-- **Learning Rate**: 2.72e-05
+- **Learning Rate**: 4.12e-05
 
 ## Model Architecture
 
@@ -54,7 +54,7 @@ This represents ongoing training of the first large-scale spiking neural network
 from huggingface_hub import hf_hub_download
 checkpoint = hf_hub_download(
     repo_id="rootxhacker/piking-llm-5b-3epochs-exp",
-    filename="checkpoint_6000.pt"
+    filename="checkpoint_8000.pt"
 )
 
 # Load with custom spiking model code
@@ -65,4 +65,4 @@ checkpoint = hf_hub_download(
 
 **🔬 This is live research in progress! Check back for updates as training continues.**
 
-**Training Progress**: 0.4% complete towards 15B tokens
+**Training Progress**: 0.5% complete towards 15B tokens

checkpoint_8000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fe8b7bcbfb9a286f983a157d1acb188863c0c9a813dc3a77c06f9c7206dace13
+size 999026318

config.json

@@ -4,11 +4,11 @@
   "hidden_size": 768,
   "num_layers": 12,
   "max_seq_length": 1024,
-  "training_step": 6000,
-  "tokens_processed": 61440000,
-  "loss": 4.738491185637147,
-  "spike_rate": 0.011329803977829647,
-  "learning_rate": 2.7196696169482876e-05,
-  "epoch": 0.012288,
-  "progress_percent": 0.40960104857868435
+  "training_step": 8000,
+  "tokens_processed": 81920000,
+  "loss": 4.727071644918862,
+  "spike_rate": 0.01129737685417216,
+  "learning_rate": 4.1215175613337367e-05,
+  "epoch": 0.016384,
+  "progress_percent": 0.5461347314382459
 }