Training checkpoint 134000 - 1.4B tokens

README.md

@@ -18,10 +18,10 @@ pipeline_tag: text-generation
 
 ## Current Training Status
 
-- **Training Step**: 132,000
-- **Tokens Processed**: 1.35B tokens
-- **Current Loss**: 4.5957
-- **Spike Rate**: 0.0503
+- **Training Step**: 134,000
+- **Tokens Processed**: 1.37B tokens
+- **Current Loss**: 4.6684
+- **Spike Rate**: 0.0504
 - **Learning Rate**: 1.97e-04
 
 ## 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_132000.pt"
+    filename="checkpoint_134000.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**: 9.0% complete towards 15B tokens
+**Training Progress**: 9.1% complete towards 15B tokens

checkpoint_134000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6c2827be20c8995355985e22f8af35c8a22d78ddea79bf285f59aa06d6b15ccb
+size 999026666

config.json

@@ -4,11 +4,11 @@
   "hidden_size": 768,
   "num_layers": 12,
   "max_seq_length": 1024,
-  "training_step": 132000,
-  "tokens_processed": 1351680000,
-  "loss": 4.5956908808605155,
-  "spike_rate": 0.05031741869168465,
-  "learning_rate": 0.00019748669378569694,
-  "epoch": 0.270336,
-  "progress_percent": 9.011223068731056
+  "training_step": 134000,
+  "tokens_processed": 1372160000,
+  "loss": 4.66839818679533,
+  "spike_rate": 0.05035190963587899,
+  "learning_rate": 0.0001973882892355905,
+  "epoch": 0.274432,
+  "progress_percent": 9.147756751590617
 }