End of training
Browse files- all_results.json +11 -11
- eval_results.json +7 -7
- runs/Oct27_22-21-52_7e534512710f/events.out.tfevents.1761658635.7e534512710f.2586.3 +3 -0
- train_results.json +5 -5
- trainer_state.json +31 -10
all_results.json
CHANGED
|
@@ -1,15 +1,15 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch": 2.
|
| 3 |
-
"eval_entropy": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_mean_token_accuracy": 0.
|
| 6 |
-
"eval_num_tokens":
|
| 7 |
-
"eval_runtime": 1046.
|
| 8 |
-
"eval_samples_per_second": 2.
|
| 9 |
"eval_steps_per_second": 0.294,
|
| 10 |
-
"total_flos": 6.
|
| 11 |
-
"train_loss": 0.
|
| 12 |
-
"train_runtime":
|
| 13 |
-
"train_samples_per_second": 7.
|
| 14 |
"train_steps_per_second": 0.15
|
| 15 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 2.501392757660167,
|
| 3 |
+
"eval_entropy": 0.43805627868353547,
|
| 4 |
+
"eval_loss": 0.4330117106437683,
|
| 5 |
+
"eval_mean_token_accuracy": 0.887996860912868,
|
| 6 |
+
"eval_num_tokens": 16158720.0,
|
| 7 |
+
"eval_runtime": 1046.2012,
|
| 8 |
+
"eval_samples_per_second": 2.352,
|
| 9 |
"eval_steps_per_second": 0.294,
|
| 10 |
+
"total_flos": 6.896237322593894e+17,
|
| 11 |
+
"train_loss": 0.04248945871988932,
|
| 12 |
+
"train_runtime": 4807.6761,
|
| 13 |
+
"train_samples_per_second": 7.166,
|
| 14 |
"train_steps_per_second": 0.15
|
| 15 |
}
|
eval_results.json
CHANGED
|
@@ -1,10 +1,10 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch": 2.
|
| 3 |
-
"eval_entropy": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_mean_token_accuracy": 0.
|
| 6 |
-
"eval_num_tokens":
|
| 7 |
-
"eval_runtime": 1046.
|
| 8 |
-
"eval_samples_per_second": 2.
|
| 9 |
"eval_steps_per_second": 0.294
|
| 10 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 2.501392757660167,
|
| 3 |
+
"eval_entropy": 0.43805627868353547,
|
| 4 |
+
"eval_loss": 0.4330117106437683,
|
| 5 |
+
"eval_mean_token_accuracy": 0.887996860912868,
|
| 6 |
+
"eval_num_tokens": 16158720.0,
|
| 7 |
+
"eval_runtime": 1046.2012,
|
| 8 |
+
"eval_samples_per_second": 2.352,
|
| 9 |
"eval_steps_per_second": 0.294
|
| 10 |
}
|
runs/Oct27_22-21-52_7e534512710f/events.out.tfevents.1761658635.7e534512710f.2586.3
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6d2587d42020e9994d2db4b9b718fd697cad9b722ed51835f797fc7f5ad3c9bc
|
| 3 |
+
size 527
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch": 2.
|
| 3 |
-
"total_flos": 6.
|
| 4 |
-
"train_loss": 0.
|
| 5 |
-
"train_runtime":
|
| 6 |
-
"train_samples_per_second": 7.
|
| 7 |
"train_steps_per_second": 0.15
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 2.501392757660167,
|
| 3 |
+
"total_flos": 6.896237322593894e+17,
|
| 4 |
+
"train_loss": 0.04248945871988932,
|
| 5 |
+
"train_runtime": 4807.6761,
|
| 6 |
+
"train_samples_per_second": 7.166,
|
| 7 |
"train_steps_per_second": 0.15
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -2,9 +2,9 @@
|
|
| 2 |
"best_global_step": 540,
|
| 3 |
"best_metric": 0.43301182985305786,
|
| 4 |
"best_model_checkpoint": "Mistral-7B-v0.1/r4/checkpoint-540",
|
| 5 |
-
"epoch": 2.
|
| 6 |
"eval_steps": 60,
|
| 7 |
-
"global_step":
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
@@ -199,12 +199,33 @@
|
|
| 199 |
"step": 540
|
| 200 |
},
|
| 201 |
{
|
| 202 |
-
"
|
| 203 |
-
"
|
| 204 |
-
"
|
| 205 |
-
"
|
| 206 |
-
"
|
| 207 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 208 |
"train_steps_per_second": 0.15
|
| 209 |
}
|
| 210 |
],
|
|
@@ -220,7 +241,7 @@
|
|
| 220 |
"early_stopping_threshold": 0.01
|
| 221 |
},
|
| 222 |
"attributes": {
|
| 223 |
-
"early_stopping_patience_counter":
|
| 224 |
}
|
| 225 |
},
|
| 226 |
"TrainerControl": {
|
|
@@ -234,7 +255,7 @@
|
|
| 234 |
"attributes": {}
|
| 235 |
}
|
| 236 |
},
|
| 237 |
-
"total_flos": 6.
|
| 238 |
"train_batch_size": 8,
|
| 239 |
"trial_name": null,
|
| 240 |
"trial_params": null
|
|
|
|
| 2 |
"best_global_step": 540,
|
| 3 |
"best_metric": 0.43301182985305786,
|
| 4 |
"best_model_checkpoint": "Mistral-7B-v0.1/r4/checkpoint-540",
|
| 5 |
+
"epoch": 2.501392757660167,
|
| 6 |
"eval_steps": 60,
|
| 7 |
+
"global_step": 600,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 199 |
"step": 540
|
| 200 |
},
|
| 201 |
{
|
| 202 |
+
"entropy": 0.4391553503357702,
|
| 203 |
+
"epoch": 2.501392757660167,
|
| 204 |
+
"grad_norm": 0.630813479423523,
|
| 205 |
+
"learning_rate": 8.903840820084096e-06,
|
| 206 |
+
"loss": 0.4249,
|
| 207 |
+
"mean_token_accuracy": 0.8895656656887796,
|
| 208 |
+
"num_tokens": 16158720.0,
|
| 209 |
+
"step": 600
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 2.501392757660167,
|
| 213 |
+
"eval_entropy": 0.4278067753880055,
|
| 214 |
+
"eval_loss": 0.4333657920360565,
|
| 215 |
+
"eval_mean_token_accuracy": 0.8881023328799706,
|
| 216 |
+
"eval_num_tokens": 16158720.0,
|
| 217 |
+
"eval_runtime": 1045.4866,
|
| 218 |
+
"eval_samples_per_second": 2.354,
|
| 219 |
+
"eval_steps_per_second": 0.295,
|
| 220 |
+
"step": 600
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 2.501392757660167,
|
| 224 |
+
"step": 600,
|
| 225 |
+
"total_flos": 6.896237322593894e+17,
|
| 226 |
+
"train_loss": 0.04248945871988932,
|
| 227 |
+
"train_runtime": 4807.6761,
|
| 228 |
+
"train_samples_per_second": 7.166,
|
| 229 |
"train_steps_per_second": 0.15
|
| 230 |
}
|
| 231 |
],
|
|
|
|
| 241 |
"early_stopping_threshold": 0.01
|
| 242 |
},
|
| 243 |
"attributes": {
|
| 244 |
+
"early_stopping_patience_counter": 9
|
| 245 |
}
|
| 246 |
},
|
| 247 |
"TrainerControl": {
|
|
|
|
| 255 |
"attributes": {}
|
| 256 |
}
|
| 257 |
},
|
| 258 |
+
"total_flos": 6.896237322593894e+17,
|
| 259 |
"train_batch_size": 8,
|
| 260 |
"trial_name": null,
|
| 261 |
"trial_params": null
|