Epoch 8. Batch Size 0. Peers 10.
Browse files- config.json +16 -5
- inner_optimizer.pt +1 -1
- model.safetensors +1 -1
- outer_optimizer.pt +1 -1
config.json
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
{
|
| 2 |
-
"_name_or_path": "
|
| 3 |
"activation_function": "gelu_new",
|
| 4 |
"all_reduce_scores": {
|
| 5 |
"0": "NON_PARTICIPATING",
|
|
@@ -89,7 +89,7 @@
|
|
| 89 |
"174": "NON_PARTICIPATING",
|
| 90 |
"175": "NON_PARTICIPATING",
|
| 91 |
"176": "NON_PARTICIPATING",
|
| 92 |
-
"177": "
|
| 93 |
"178": "NON_PARTICIPATING",
|
| 94 |
"179": "NON_PARTICIPATING",
|
| 95 |
"18": "NON_PARTICIPATING",
|
|
@@ -148,7 +148,7 @@
|
|
| 148 |
"227": "NON_PARTICIPATING",
|
| 149 |
"228": "NON_PARTICIPATING",
|
| 150 |
"229": "NON_PARTICIPATING",
|
| 151 |
-
"23": "
|
| 152 |
"230": "NON_PARTICIPATING",
|
| 153 |
"231": "NON_PARTICIPATING",
|
| 154 |
"232": "NON_PARTICIPATING",
|
|
@@ -267,7 +267,18 @@
|
|
| 267 |
"AutoConfig": "distributed/optimized-gpt2-500m--configuration_gpt_optimized.GPTOptimConfig",
|
| 268 |
"AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
|
| 269 |
},
|
| 270 |
-
"block_list": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 271 |
"block_size": 1024,
|
| 272 |
"bos_token_id": 50256,
|
| 273 |
"embd_pdrop": 0.1,
|
|
@@ -275,7 +286,7 @@
|
|
| 275 |
"initializer_range": 0.02,
|
| 276 |
"inner_step": 0,
|
| 277 |
"inner_steps": 0,
|
| 278 |
-
"last_allreduce_block":
|
| 279 |
"layer_norm_epsilon": 1e-05,
|
| 280 |
"model_type": "gpt_optimized",
|
| 281 |
"n_embd": 1280,
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "kmfoda/gpt2-1b-miner-3",
|
| 3 |
"activation_function": "gelu_new",
|
| 4 |
"all_reduce_scores": {
|
| 5 |
"0": "NON_PARTICIPATING",
|
|
|
|
| 89 |
"174": "NON_PARTICIPATING",
|
| 90 |
"175": "NON_PARTICIPATING",
|
| 91 |
"176": "NON_PARTICIPATING",
|
| 92 |
+
"177": "NON_PARTICIPATING",
|
| 93 |
"178": "NON_PARTICIPATING",
|
| 94 |
"179": "NON_PARTICIPATING",
|
| 95 |
"18": "NON_PARTICIPATING",
|
|
|
|
| 148 |
"227": "NON_PARTICIPATING",
|
| 149 |
"228": "NON_PARTICIPATING",
|
| 150 |
"229": "NON_PARTICIPATING",
|
| 151 |
+
"23": "NON_PARTICIPATING",
|
| 152 |
"230": "NON_PARTICIPATING",
|
| 153 |
"231": "NON_PARTICIPATING",
|
| 154 |
"232": "NON_PARTICIPATING",
|
|
|
|
| 267 |
"AutoConfig": "distributed/optimized-gpt2-500m--configuration_gpt_optimized.GPTOptimConfig",
|
| 268 |
"AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
|
| 269 |
},
|
| 270 |
+
"block_list": [
|
| 271 |
+
5283473,
|
| 272 |
+
5283477,
|
| 273 |
+
5283481,
|
| 274 |
+
5283486,
|
| 275 |
+
5283490,
|
| 276 |
+
5283494,
|
| 277 |
+
5283499,
|
| 278 |
+
5283503,
|
| 279 |
+
5283508,
|
| 280 |
+
5283512
|
| 281 |
+
],
|
| 282 |
"block_size": 1024,
|
| 283 |
"bos_token_id": 50256,
|
| 284 |
"embd_pdrop": 0.1,
|
|
|
|
| 286 |
"initializer_range": 0.02,
|
| 287 |
"inner_step": 0,
|
| 288 |
"inner_steps": 0,
|
| 289 |
+
"last_allreduce_block": 5283755,
|
| 290 |
"layer_norm_epsilon": 1e-05,
|
| 291 |
"model_type": "gpt_optimized",
|
| 292 |
"n_embd": 1280,
|
inner_optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 2752
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:12806c5436a164b1e90e09a1518f6a88b15add872d3c4f6abf840225a7ecb68a
|
| 3 |
size 2752
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4040701744
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c933a9ed1f5f7c6efadfcb8d1009c01cfbc8fb83615528124ae7a9849a72cf46
|
| 3 |
size 4040701744
|
outer_optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4040805354
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:acccf246a3b110692024e06f5a90c3a584dcf61fbc1167125b6943a2a1aa3f8d
|
| 3 |
size 4040805354
|