Training checkpoint 176000 - 1.8B tokens

README.md

@@ -18,9 +18,9 @@ pipeline_tag: text-generation
 
 ## Current Training Status
 
-- **Training Step**: 174,000
-- **Tokens Processed**: 1.78B tokens
-- **Current Loss**: 4.6834
+- **Training Step**: 176,000
+- **Tokens Processed**: 1.80B tokens
+- **Current Loss**: 4.6899
 - **Spike Rate**: 0.0504
 - **Learning Rate**: 1.95e-04
 
@@ -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_174000.pt"
+    filename="checkpoint_176000.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**: 11.9% complete towards 15B tokens
+**Training Progress**: 12.0% complete towards 15B tokens

checkpoint_176000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:69c7d7603181c47cbe4a1ce3f560600867c69910ba0a89ddc8034330911cdf0b
+size 999026666

config.json

@@ -4,11 +4,11 @@
   "hidden_size": 768,
   "num_layers": 12,
   "max_seq_length": 1024,
-  "training_step": 174000,
-  "tokens_processed": 1781760000,
-  "loss": 4.683350258383383,
-  "spike_rate": 0.05041792342778572,
-  "learning_rate": 0.0001950314982278183,
-  "epoch": 0.356352,
-  "progress_percent": 11.878430408781847
+  "training_step": 176000,
+  "tokens_processed": 1802240000,
+  "loss": 4.68992041871115,
+  "spike_rate": 0.05041941014635301,
+  "learning_rate": 0.00019489439751956046,
+  "epoch": 0.360448,
+  "progress_percent": 12.014964091641408
 }